System zum patientenspezifischen Modellieren von Blutfluss

System zum Bestimmen von kardiovaskulären Informationen für einen Patienten, wobei das System Folgendes umfasst: wenigstens ein Computersystem, das konfiguriert ist, um: patientenspezifische Daten bezüglich einer Geometrie einer anatomischen Struktur des Patienten zu empfangen, wobei die anatomische Struktur wenigstens einen Abschnitt einer Mehrzahl an Koronararterien, die von einer Aorta ausgehen, beinhaltet; basierend auf den patientenspezifischen Daten ein dreidimensionales Modell zu erzeugen, das einen ersten Abschnitt der anatomischen Struktur repräsentiert, wobei der erste Abschnitt der anatomischen Struktur wenigstens den Abschnitt der Mehrzahl an Koronararterien beinhaltet; wenigstens teilweise basierend auf einer Masse oder einem Volumen des Myokardgewebes ein physikbasiertes Modell bezüglich einer Blutflusseigenschaft im ersten Abschnitt der anatomischen Struktur zu erzeugen; und wenigstens teilweise basierend auf dem dreidimensionalen Modell und dem physikbasierten Modell eine ...

PROCEDURE FOR THE QUANTIFICATION OF AN AT THE BASIS LYING CHARACTERISTICS OF A QUANTITY OF SAMPLES

DATA MINING BASED STUDENT DATA ANALYSIS UTILIZING RANDOM FOREST METHOD

DATA MINING BASED STUDENT DATA ANALYSIS UTILIZING RANDOM FOREST METHOD Students are the strength of any educational institution whose quality is measured by the process of official recognition termed as Accreditation. The measurement of quality focuses on strength of students hence educational institutions give more importance to prevent drop out of students as high dropping rate of students create bad impact on the institutions resulting in low grade accreditation and finally ends in bad reputation. This invention involves data mining technique for analyzing the student educational data which is utilized by the classification method for predicting the student drop out in undergraduate level at XYZ University. Prediction of student drop out is done based on the raw student academics data enrolled in the university for a particular academic year. Imbalanced raw data is handled through preprocessing where in our invention imbalance dataset is handled by synthetic minority oversampling technique ...

Method and system for patient-specific modeling of blood flow

C:\Users\kll\AppData\Local\Temp12713 IDAOBAFC.DOCX-22 12/2015 Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

Method and system for patient-specific modeling of blood flow

Abstract A system for determining cardiovascular information for a patient, the system comprising: at least one computer system configured to: receive patient specific data regarding a geometry of an anatomical structure of the patient; create a three-dimensional model representing at least a portion of the anatomical structure of the patient based on the patient-specific data, the three-dimensional model representing at least one fluid flow inlet and at least one fluid flow outlet; create at least one boundary condition model representing fluid flow through at least one of the at least one inlet or the at least one outlet, based at least in part on modeling a condition of hyperemia; and determine first information regarding a blood flow characteristic within the anatomical structure of the patient based on the three dimensional model and the at least one boundary condition model. cCD Co co VC C C) C) f I co C, Loij 0~ u~ EtIt C3oI ...

Distributed hierarchical evolutionary modeling and visualization of empirical data

Image recognition method

An image recognition method includes the steps of: obtaining a pixel value distribution of a plurality of pixels within a first selected block in the last M images of a first monitoring video stream as a first interval pixel distribution, obtaining a pixel value distribution of a plurality of pixels within a first selected block in the last N images of a first monitoring video stream as a second interval pixel distribution, obtaining a first variance coefficient corresponding to the first selected block based on the first interval pixel distribution and the second interval pixel distribution, and generating a first recognition signal when the first variance coefficient is greater than a first threshold.

FEATURE VALUE CANDIDATE GENERATING DEVICE AND FEATURE VALUE CANDIDATE GENERATING METHOD

A feature value candidate generating device comprises storage means storing feature values of plural types extracted from samples, index value calculating means for calculating the index values obtained by normalizing the number of feature values with the number of samples for each feature value, an evaluation subject selecting means for selecting a combination of feature values to be evaluated from the feature values of the plural types, evaluating means for evaluating the combination of the selected feature values as evaluation subjects by judging whether or not the uniformity of the frequency distribution of the index values of the feature values meets a predetermined criterion, and candidate determining means for determining the combination of the feature values evaluated as meeting the predetermined criterion as feature value candidates given to a model making device.

Method for pan-sharpening panchromatic and multispectral images using dictionaries

A single panchromatic (Pan) image and a single multispectral (MS) image are Pan-sharpened by extracting features from the Pan image and the MS image. The features are decomposed into features without missing values and features with missing values. A dictionary is learned from the features without missing values. The values for the features with the missing values are learned using the dictionary. The MS image is merged with the Pan image including the predicted values into a merged image, and the merged image is then Pan sharpened.

System and method for detecting anomalies in images

Anomalies in images are detected. A generative network and/or an autoencoder (“G/A-Network”), a Siamese network, a first training-dataset of normal images and a second training-dataset of abnormal images are provided. The G/A-network is trained to produce latent data from input images and output images from the latent data, wherein the training is performed with images of the first training-dataset, wherein a loss function is used for training at least at the beginning of training, and the loss function enhances the similarity of the input images and respective output images. The Siamese network is trained to generate similarity measures between input images and respective output images, wherein the training is performed with images of the first training-dataset and the second training-dataset in that images of both training-datasets are used as input images for the G/A-network and output images of the G/A-network are compared with their respective input images by the Siamese network. 1. A method for producing a system for detecting anomalies in images , the method comprising:providing a generative network and/or an autoencoder,providing a Siamese network,providing a first training-dataset comprising normal images,providing a second training-dataset comprising abnormal images,training the generative network and/or the autoencoder to produce latent data from input images and output images from the latent data, wherein the training is performed with images of the first training-dataset, wherein a loss function is used for training at least at the beginning of training, the loss function enhancing a similarity of the input images and respective output images, andtraining the Siamese network to generate similarity measures between the input images and the respective output images, wherein the training is performed with images of the first training-dataset and the second training-dataset in that images of both training-datasets are used as input images for the generative ...

Multi-task multi-modal machine learning system

Methods, systems, and apparatus, including computer programs encoded on computer storage media for training a machine learning model to perform multiple machine learning tasks from multiple machine learning domains. One system includes a machine learning model that includes multiple input modality neural networks corresponding to respective different modalities and being configured to map received data inputs of the corresponding modality to mapped data inputs from a unified representation space; an encoder neural network configured to process mapped data inputs from the unified representation space to generate respective encoder data outputs; a decoder neural network configured to process encoder data outputs to generate respective decoder data outputs from the unified representation space; and multiple output modality neural networks corresponding to respective different modalities and being configured to map decoder data outputs to data outputs of the corresponding modality.

IMAGE PROCESSING AND PATIENT-SPECIFIC MODELING OF BLOOD FLOW

Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three-dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

POWER SAVING TECHNIQUES FOR AN IMAGE CAPTURE DEVICE

An image capture device that includes an adjustment circuit configured to monitor image parameters, generate updated image settings for the image capture device in response to the monitored image parameters, and transmit the updated image settings to one or more processors. The updated image settings configure the one or more processors to determine whether to transition the image capture device from a dynamic scene mode to a static scene mode based on a first image parameter included in the monitored image parameters, wherein the first image parameter is different from a second image parameter used to determine to transition the image capture device from the static scene mode to the dynamic scene mode, and to suspend generation of all or less than all of the updated image settings in response to determining to transition the image capture device from the dynamic scene mode to the static scene mode. 1. An image capture device , comprising:an adjustment circuit configured to monitor image parameters, generate updated image settings for the image capture device based on the monitored image parameters, and transmit the updated image settings; andone or more processors configured to receive the transmitted updated image settings from the adjustment circuit, wherein the received updated image settings comprise instructions for configuring the one or more processors to perform image processing of the image capture device, and wherein the updated image settings configure the one or more processors to:determine to transition the image capture device from a dynamic scene mode to a static scene mode based on a first image parameter included in the monitored image parameters, wherein the first image parameter used to determine to transition the image capture device from the dynamic scene mode to the static scene mode is different from a second image parameter that the one or more processors use to determine to transition the image capture device from the static scene mode to ...

EYE STATE DETECTION METHOD, ELECTRONIC DEVICE, DETECTING APPARATUS AND COMPUTER READABLE STORAGE MEDIUM

An eye state detection method, including: acquiring a target image (101); positioning a plurality of eye feature points in the target image to determine position coordinates of the plurality of eye feature points (102); normalizing the position coordinates of the plurality of eye feature points to obtain normalized position feature data (103); and determining an eye state in the target image based on the position feature data (104). An eye state detection apparatus, a device and a medium are also provided.

Systems and methods for automatic scale selection in real-time imaging

A system and method for automatic scale selection in real-time image and video processing and computer vision applications. In one aspect, a non-parametric variable bandwidth mean shift technique, which is based on adaptive estimation of a normalized density gradient, is used for detecting one or more modes in the underlying data and clustering the underlying data. In another aspect, a data-driven bandwidth (or scale) selection technique is provided for the variable bandwidth mean shift method, which estimates for each data point the covariance matrix that is the most stable across a plurality of scales. The methods can be used for detecting modes and clustering data for various types of data such as image data, video data speech data, handwriting data, etc.

Systems and Methods for Predicting Instance Geometry

Systems and methods for predicting instance geometry are provided. A method includes obtaining an input image depicting at least one object. The method includes determining an instance mask for the object by inputting the input image into a machine-learned instance segmentation model. The method includes determining an initial polygon with a number of initial vertices outlining the border of the object within the input image. The method includes obtaining a feature embedding for one or more pixels of the input image and determining a vertex embedding including a feature embedding for each pixel corresponding an initial vertex of the initial polygon. The method includes determining a vertex offset for each initial vertex of the initial polygon based on the vertex embedding and applying the vertex offset to the initial polygon to obtain one or more enhanced polygons.

SYSTEMS, METHODS, AND DEVICES FOR MEDICAL IMAGE ANALYSIS, DIAGNOSIS, RISK STRATIFICATION, DECISION MAKING AND/OR DISEASE TRACKING

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters.

Method and system for patient-specific modeling of blood flow

C:\Users\kIl\AppData\Locl\Temp12713_1 DAOBAFC.DOCX-22 12 2015 Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

Method and system for patient-specific modeling of blood flow

C:\Users\kll\AppData\Local\Temp12713 IDAOBAFC.DOCX-22 12/2015 Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

DISTRIBUTED HIERARCHICAL EVOLUTIONARY MODELING AND VISUALIZATION OF EMPIRICAL DATA

A distributed hierarchical evolutionary modeling and visualization of empirical data method and machine readable storage medium for creating an empirical modeling system based upon previously acquired data. The data represents inputs to the systems and corresponding outputs from the system. The method and machine readable storage medium utilize an entropy fonction based upon information theory and the principles of thermodynamics to accurately predict system outputs from subsequently acquired inputs. The method and machine readable storage medium identify the most information-rich (i.e., optimum) representation of a data set in order to reveal the underlying order, or structure, of what appears to be a disordered system. Evolutionary programming is one method utilized for identifying the optimum representation of data.

SYSTEMS, METHODS AND DEVICES FOR MONITORING BETTING ACTIVITIES

System, processes and devices for monitoring betting activities using bet recognition devices and a server. Each bet recognition device has an imaging component for capturing image data for a gaming table surface The bet recognition device receives calibration data for calibrating the bet recognition device. A server processor coupled to a data store processes the image data received from the bet recognition devices over the network to detect, for each betting area, a number of chips and a final bet value for the chips.

SYSTEMS, METHODS AND DEVICES FOR MONITORING BETTING ACTIVITIES

System, processes and devices for monitoring betting activities using bet recognition devices and a server. Each bet recognition device has an imaging component for capturing image data for a gaming table surface The bet recognition device receives calibration data for calibrating the bet recognition device. A server processor coupled to a data store processes the image data received from the bet recognition devices over the network to detect, for each betting area, a number of chips and a final bet value for the chips.

METHOD AND SYSTEM FOR PATIENT-SPECIFIC MODELING OF BLOOD FLOW

QUANTITATIVE DNA-BASED IMAGING AND SUPER-RESOLUTION IMAGING

The present disclosure provides, inter alia, methods and compositions (e.g., conjugates) for imaging, at high spatial resolution, targets of interest.

Method and system for patient-specific modeling of blood flow

ACCURATE ROI EXTRACTION METHOD AND SYSTEM AIDED BY OBJECT TRACKING

Modeling method of automatic valuation model

Unmanned aerial vehicle reconnaissance target evolution rule prediction method based on circular neural network

Independent component analysis of tensor for sensor data fusion and reconstruction

특징 추출 방법 및 장치

... 본 발명은 특징 추출 방법 및 장치를 개시하고, 영상 처리 기술 분야에 속한다. 상기 특징 추출 방법은, 영상을 복수 개의 셀을 포함하는 블록으로 구분하는 단계; 각 셀을 공간 영역에서 주파수 영역으로 변환하는 단계; 영상의 주파수 영역에서의 방향 그라데이션 히스토그램(HOG) 특징을 추출하는 단계를 포함한다. 주파수 영역에서 영상의 HOG 특징을 추출함으로써 HOG 특징을 추출하는 과정에서 영상의 공간 영역에 대해 직접 계산하여 얻음으로 인한 패턴 인식에서의 검출율과 정확도가 낮은 문제점를 해결하고; 주파수 영역에서 영상의 HOG 특징을 추출함으로써, 패턴 인식에서의 검출율과 정확도를 향상시키는 효과를 달성할 수 있다.

환자별 혈류 모델링 방법 및 시스템

... 본 발명의 환자의 심혈관 정보를 결정하기 위한 시스템에 관한 것으로, 상기 시스템은 적어도 하나의 컴퓨터 시스템을 포함하며, 상기 적어도 하나의 컴퓨터 시스템은, 환자 심장의 기하 형태에 관한 환자별 데이터를 수신하도록 구성되고, 환자별 데이터에 기초하여 환자 심장의 적어도 일부분을 나타내는 3차원 모델을 생성하도록 구성되며, 환자 심장의 혈류 특성에 관한 물리학-기반 모델을 생성하도록 구성되고, 상기 3차원 모델 및 물리학-기반 모델에 기초하여 상기 환자 심장 내의 분획 혈류 예비력(FFR: Fractional Flow Reserve)을 결정하도록 구성된다.

Machine learning by construction of a decision function

Data processing apparatus is provided for evaluating answers to respective query items considered to be represented by respective points within a region of feature space, and the apparatus comprises an input (10) which receives such a query item. The region is subdivided into subregions according to at least first and second subdivisions. A subregion identifying portion (20) identifies, for each such subdivision of the region, which subregion of the subdivision contains the point representing the received query item. A partial answer retrieval portion (30) has access when the apparatus is in use to a store (40) of precalculated partial answers for at least some the subregions of the subdivisions, and retrieves from the store the partial answers for the or each identified subregion that is present in the store. An answer calculation portion (50) calculates an answer to the received query item based on the retrieved partial answers, and an output (60) outputs the calculated answer.

Method for detecting the possible taking of screenshots

A computer-implemented method for sending a message from a first mobile device to a second mobile device comprises transmitting the message by the first mobile device, receiving the message by the second mobile device, displaying the message on the second mobile device, activating a camera located on the second mobile device in order to capture image data regarding a field of view of the camera, monitoring the image data to detect a possible presence of one or more suspected camera lenses within the field of view of the camera, and upon detection of the possible presence of the one or more suspected camera lenses within the field of view of the camera, ceasing displaying the message on the second mobile device.

PROCESSING APPARATUS, PROCESSING METHOD, LEARNING APPARATUS, AND COMPUTER PROGRAM PRODUCT

According to an embodiment, a processing apparatus includes a hardware processor. The hardware processor is configured to: cut out, from an input signal, a plurality of partial signals that are predetermined parts in the input signal; execute processing on the plurality of partial signals using neural networks having the same layer structure with each other to generate a plurality of intermediate signals including a plurality of signals corresponding to a plurality of channels; execute predetermined statistical processing on signals for each of the plurality of channels for each of the plurality of intermediate signals corresponding to the plurality of partial signals, to calculate statistics for each channel and generate a concatenated signal by concatenating the statistics of the plurality of respective intermediate signals for each channel; generate a synthetic signal by performing predetermined processing on the concatenated signal; and output an output signal in accordance with the ...

QUANTITATIVE DNA-BASED IMAGING AND SUPER-RESOLUTION IMAGING

The present disclosure provides, inter alia, methods and compositions (e.g., conjugates) for imaging, at high spatial resolution, targets of interest. 1. A protein-nucleic acid conjugate , comprising a protein linked to a docking strand that is capable of transiently binding to a complementary labeled imager strand.234.-. (canceled)35. A method of detecting a target in a sample , the method comprising:contacting the sample with (a) at least one antibody-DNA conjugate that comprises a biotinylated antibody linked to a biotinylated docking strand through a biotin-streptavidin linker, and (b) at least one fluorescently-labeled imager strand that is complementary to the docking strand of the at least one antibody-DNA conjugate; anddetermining whether the at least one antibody-DNA conjugate binds to the target in the sample.36. The method of claim 35 , wherein the determining step comprises imaging binding of the at least one fluorescently-labeled imager strand to the docking strand of the at least one antibody-DNA conjugate.37. (canceled)38. The method of claim 35 , wherein the antibody is a monoclonal antibody.3946.-. (canceled)47. The method of claim 35 , wherein the docking strand comprises at least two domains claim 35 , wherein each domain is complementary to a respectively labeled imager strand.4849.-. (canceled)50. The method of claim 1 , comprising:contacting the sample with (a) at least two different antibody-DNA conjugates, each comprising a biotinylated antibody linked to a biotinylated docking strand through an intermediate biotin-streptavidin linker, and (b) at least two labeled imager strands that are complementary to respective docking strands of the at least two different antibody-DNA conjugates; anddetermining whether the at least two antibody-DNA conjugates bind to at least two targets in the sample.5168.-. (canceled)69. The method of claim 50 , wherein the sample is contacted sequentially with the at least two labeled imager strands of (b).70. The ...

SOFTWARE COMPONENT DEFECT PREDICTION USING CLASSIFICATION MODELS THAT GENERATE HIERARCHICAL COMPONENT CLASSIFICATIONS

Systems and methods for facilitating updates to software programs via machine-learning techniques are disclosed. In an example, an application generates a feature vector from a textual description of a software defect by applying a topic model to the textual description. The application uses the feature vector and one or more machine-learning models configured to predict classifications and sub-classifications of the textual description. The application integrates the classifications and the sub-classifications into a final classification of the textual description that indicates a software component responsible for causing the software defect. The final classification is usable for correcting the software defect.

Human monitoring system incorporating calibration methodology

Related methods are provided for establishing a baseline value to represent an eyelid opening dimension for a person engaged in an activity, where the activity may be driving a vehicle, operating industrial equipment, or performing a monitoring or control function; and for operating a system for monitoring eyelid opening values with real time video data.

Method for quantifying an underlying property of a multitude of samples

An approach for quantification of an underlying property of a multitude of samples, e. g. cellular biological samples, is described. In an exemplary application, input measures are mean pixel intensities from different segments of individual cells on original and filtered images. In another exemplary application, input measures are pixel intensities from different cells on a series of time-gated images. The set of input measures are expected to have a linear relationship to the underlying property. Latent variables are (1) a scalar noise factor responsible for intensity fluctuations in all segments of a cell at a time and (2) the main underlying property of the process responsible for observed changes on images. Linear coefficients for latent variables are determined using noise minimization while constraining the mean signal from positive and negative control samples to the corresponding theoretical values. The method is well suited for high throughput applications like drug screening.

СПОСОБ ОПРЕДЕЛЕНИЯ МЕСТОНАХОЖДЕНИЯ ЛЕЙКОЦИТОВ КОСТНОГО МОЗГА НА ОСНОВЕ АГРЕГАЦИИ НАСЫЩЕНИЯ

Изобретение относится к области обработки медицинских изображений. Техническим результатом является повышение точности определения местоположения лейкоцитов костного мозга на изображении. Способ определения местонахождения лейкоцитов костного мозга включает следующие этапы: осуществляют медианную фильтрацию изображения лейкоцитов костного мозга; производят изменение цвета изображения лейкоцитов костного мозга; применяют алгоритм поиска K-средних к каналу S-насыщения; производят вычисление среднего значения первых двух частей канала H и вычисление первых двух частей в соответствии со средними точками первых двух частей; рассчитывают составные части лейкоцитов на нескольких изображениях и записывают значения; в соответствии с записанными результатами применяют алгоритм дерева решений; осуществляют морфологическую обработку результатов предыдущего этапа; устанавливают расположения лейкоцитов, выделенных на предыдущем этапе. 3 з.п. ф-лы, 7 ил.

Model generation having account of variability

A computer-implemented method of generating a model from a set of images. The method comprises processing a plurality of data items, each data item representing an image of said set of images, to determine variability between said plurality of data items; and generating model data representing said model based upon said data items and said variability, wherein the influence of each of said data items upon the generated model is determined by a relationship between a respective one of said data items and said variability. The invention may be applicable to detection of defects for instance during industrial image inspection.

System for distributed data processing using clustering

Neural network training data selection using memory reduced cluster analysis for field model development

NEURAL NETWORK TRAINING DATA SELECTION USING MEMORY REDUCED CLUSTER ANALYSIS FOR FIELD MODEL DEVELOPMENT

Method and system for patient-specific modeling of blood flow

C:\Users\kIl\AppData\Locl\Temp12713_1 DAOBAFC.DOCX-22 12 2015 Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

Image searching method and apparatus

Apparatus for performing searching of a plurality of reference images, the apparatus including one or more electronic processing devices that search the plurality of reference images to identify first reference images similar to a sample image, identify image tags associated with at least one of the first reference image, search the plurality of reference images to identify second reference images using at least one of the image tags and provide search results including at least some first and second reference images.

Batch normalization layers

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer.

BATCH NORMALIZATION LAYERS

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer. Fig. 1. WO 2016/123409 PCT/US2016/015476 Neural Network Outputs Neural Network System 100 Neural Network Layer B112 Batch Normalization Layer Outputs 110 Batch Normalization Layer 108 Layer A Outputs 106 ...

환자별 혈류 모델링 방법 및 시스템

... 본 발명의 환자의 심혈관 정보를 결정하기 위한 시스템에 관한 것으로, 상기 시스템은 적어도 하나의 컴퓨터 시스템을 포함하며, 상기 적어도 하나의 컴퓨터 시스템은, 환자 심장의 기하 형태에 관한 환자별 데이터를 수신하도록 구성되고, 환자별 데이터에 기초하여 환자 심장의 적어도 일부분을 나타내는 3차원 모델을 생성하도록 구성되며, 환자 심장의 혈류 특성에 관한 물리학-기반 모델을 생성하도록 구성되고, 상기 3차원 모델 및 물리학-기반 모델에 기초하여 상기 환자 심장 내의 분획 혈류 예비력(FFR: Fractional Flow Reserve)을 결정하도록 구성된다.

PAIN MONITORING USING MULTIDIMENSIONAL ANALYSIS OF PHYSIOLOGICAL SIGNALS

The invention discloses a method and system for establishing the pain level in an awake, semi-awake or sedated patient. The method comprises steps of analyzing a multidimensional array of physiological signals to obtain the pain level of a patient. The signals are processed so as to extract a vector of Great Plurality of Features representing the patient's physiological status. The vector of the Great Plurality of Features is processed and classified into at least two classes for at least two conditions. These classes represent the pain level of the patient at a given time interval and thereby are used to establish the pain level of an awake, semi- awake or sedated patient. A system is provided for establishing the pain level in an awake, semi- awake or sedated patient.

ACCURACY OF STREAMING RNN TRANSDUCER

A computer-implemented method is provided for model training. The method includes training a second end-to-end neural speech recognition model that has a bidirectional encoder to output same symbols from an output probability lattice of the second end-to-end neural speech recognition model as from an output probability lattice of a trained first end-to-end neural speech recognition model having a unidirectional encoder. The method also includes building a third end-to-end neural speech recognition model that has a unidirectional encoder by training the third end-to-end neural speech recognition model as a student by using the trained second end-to-end neural speech recognition model as a teacher in a knowledge distillation method.

Anomaly detection method, program, and system

A method providing an analytical technique introducing label information into an anomaly detection model. The method includes the steps of: inputting measurement data having an anomalous or normal label and measurement data having no label as samples; determining a similarity matrix indicating the relationship between the samples based on the samples; defining a penalty based on the similarity matrix and calculating parameters in accordance with an updating equation having a term reducing the penalty; and calculating a degree of anomaly based on the calculated parameters. The present invention also provides a program and system for detecting an anomaly based on measurement data.

Image processing method and image processing system

An image processing method applied to an image processing system. The image processing method comprises: (a) computing an image intensity distribution of an input image; (b) performing atmospheric light estimation to the input image; (c) performing transmission estimation according to a result of the step (a) to the input image, to generate a transmission estimation parameter; and (d) recovering scene radiance of the input image according to a result generated by the step (b) and the transmission estimation parameter. At least one of the steps (a)-(c) are performed to data corresponding to only partial pixels of the input image.

Power saving techniques for an image capture device

An image capture device that includes an adjustment circuit configured to monitor image parameters, generate updated image settings for the image capture device in response to the monitored image parameters, and transmit the updated image settings to one or more processors. The updated image settings configure the one or more processors to determine whether to transition the image capture device from a dynamic scene mode to a static scene mode based on a first image parameter included in the monitored image parameters, wherein the first image parameter is different from a second image parameter used to determine to transition the image capture device from the static scene mode to the dynamic scene mode, and to suspend generation of all or less than all of the updated image settings in response to determining to transition the image capture device from the dynamic scene mode to the static scene mode.

Method and apparatus for image processing

A method is provided for recognition of a ceiling portion, a vertical object portion and a ground portion in an image of indoor scene executed in an electronic system. The image is divided into a plurality of pixel sets. Expected values of each pixel sets with a ceiling distribution function, a vertical object distribution function and a ground distribution function are calculated. The expected values of each pixel set in the ceiling distribution function, the vertical object distribution function and the ground distribution function are compared to determine whether each pixel set belongs to a ceiling object, a vertical object or a ground object.

Systems and methods for automatic scale selection in real-time imaging

A system and method for automatic scale selection in real-time image and video processing and computer vision applications. In one aspect, a non-parametric variable bandwidth mean shift technique, which is based on adaptive estimation of a normalized density gradient, is used for detecting one or more modes in the underlying data and clustering the underlying data. In another aspect, a data-driven bandwidth (or scale) selection technique is provided for the variable bandwidth mean shift method, which estimates for each data point the covariance matrix that is the most stable across a plurality of scales. The methods can be used for detecting modes and clustering data for various types of data such as image data, video data speech data, handwriting data, etc.

Anomaly detection from aggregate statistics using neural networks

Implementations disclosed describe a method and a system to perform the method of obtaining a reduced representation of a plurality of sensor statistics representative of data collected by a plurality of sensors associated with a device manufacturing system performing a manufacturing operation. The method further includes generating, using a plurality of outlier detection models, a plurality of outlier scores, each of the plurality of outlier scores generated based on the reduced representation of the plurality of sensor statistics using a respective one of the plurality of outlier detection models. The method further includes processing the plurality of outlier scores using a detector neural network to generate an anomaly score indicative of a likelihood of an anomaly associated with the manufacturing operation.

Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters.

CLUSTER TECHNOLOGY FOR CYCLIC PHENOMENA

QR code decoding system

Method and system for patient-specific modeling of blood flow

C:\Users\kll\AppData\Local\Temp12713 IDAOBAFC.DOCX-22 12/2015 Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

REFINED AVERAGE FOR ZONING METHOD AND SYSTEM

A method for determining management zones within an agricultural field, the method includes selecting a plurality of remotely sensed images of the agricultural field wherein the plurality of remotely sensed images represent a plurality of growing seasons, each of the plurality of remotely sensed images having a vegetation index associated therewith, generating a refined average image from the plurality of remotely sensed images of the agricultural field, and applying a classification method to define management zones associated with the refined average image.

RADIO-WAVE DETECTION DEVICE

The present invention is provided with: a prediction unit (19) that predicts a feature amount which can be extracted in the future by a feature amount extraction unit (12) by use of a prediction model for the feature amount; and a hypothesis generation unit (15) that, by using a plurality of feature amounts extracted by the feature amount extraction unit (12) and the feature amount predicted by the prediction unit (19), generates a hypothesis for hypothesizing about a transmission source of a radio wave detected by a reception unit (11).

SYSTEMS, METHODS AND DEVICES FOR MONITORING BETTING ACTIVITIES

System, processes and devices for monitoring betting activities using bet recognition devices and a server. Each bet recognition device has an imaging component for capturing image data for a gaming table surface The bet recognition device receives calibration data for calibrating the bet recognition device. A server processor coupled to a data store processes the image data received from the bet recognition devices over the network to detect, for each betting area, a number of chips and a final bet value for the chips.

DISTRIBUTED HIERARCHICAL EVOLUTIONARY MODELING AND VISUALIZATION OF EMPIRICAL DATA

A distributed hierarchical evolutionary modeling and visualization of empirical data method and machine readable storage medium for creating an empirical modeling system based upon previously acquired data. The data represents inputs to the systems and corresponding outputs from the system. The method and machine readable storage medium utilize an entropy fonction based upon information theory and the principles of thermodynamics to accurately predict system outputs from subsequently acquired inputs. The method and machine readable storage medium identify the most information-rich (i.e., optimum) representation of a data set in order to reveal the underlying order, or structure, of what appears to be a disordered system. Evolutionary programming is one method utilized for identifying the optimum representation of data.

환자별 혈류 모델링 방법 및 시스템

... 본 발명의 환자의 심혈관 정보를 결정하기 위한 시스템에 관한 것으로, 상기 시스템은 적어도 하나의 컴퓨터 시스템을 포함하며, 상기 적어도 하나의 컴퓨터 시스템은, 환자 심장의 기하 형태에 관한 환자별 데이터를 수신하도록 구성되고, 환자별 데이터에 기초하여 환자 심장의 적어도 일부분을 나타내는 3차원 모델을 생성하도록 구성되며, 환자 심장의 혈류 특성에 관한 물리학-기반 모델을 생성하도록 구성되고, 상기 3차원 모델 및 물리학-기반 모델에 기초하여 상기 환자 심장 내의 분획 혈류 예비력(FFR: Fractional Flow Reserve)을 결정하도록 구성된다.

SYSTEM AND METHOD FOR EVALUATING DOCUMENT CLASSIFICATION USING APPROPRIATENESS OF CLASS MODELS, PATTERN CLASSIFICATION EVALUATING PROGRAM, AND STORING MEDIUM OR DEVICE

PURPOSE: A system and a method for evaluating document classification using appropriateness of class models, a pattern classification evaluating program, and a storing medium or device are provided to reduce burden for designing a document classification system and reconfiguring the class models by facilitating detection of a class pair approximated to a topic and a class of the deteriorated class model. CONSTITUTION: A document input block(210) receives a document set. A document preprocessing block(220) performs term extraction, morphological analysis, and document vector formation for the inputted document. A training document information storing block(240) stores training document information for each previously prepared class. An actual document information storing block(250) stores actual document information for each class obtained based on a classification result. A document information processing block(230) calculates similarity among all class pairs for a training document set ...

LAND USAGE PROPERTY IDENTIFICATION METHOD, APPARATUS, ELECTRONIC DEVICE AND STORAGE MEDIUM

A land usage property identification method, apparatus, electronic device and storage medium are disclosed. The method includes: acquiring point-of-interest (POI) data and area-of-interest (AOI) data; dividing a target area to be identified according to road network information, and obtaining at least one block in the target area; associating acquired POI data to a corresponding target block in the at least one block; obtaining a first weight set corresponding to a corresponding category of each POI data in the target block; obtaining a second weight set corresponding to a corresponding area of each AOI data in the target block; obtaining a land usage property weight set according to the first weight set, the second weight set and a preset land usage classification standard; and identifying a land usage property of the target block according to a target weight in the land usage property weight set.

Fully parallel, low complexity approach to solving computer vision problems

Values of pixels in an image are mapped to a binary space using a first function that preserves characteristics of values of the pixels. Labels are iteratively assigned to the pixels in the image in parallel based on a second function. The label assigned to each pixel is determined based on values of a set of nearest-neighbor pixels. The first function is trained to map values of pixels in a set of training images to the binary space and the second function is trained to assign labels to the pixels in the set of training images. Considering only the nearest neighbors in the inference scheme results in a computational complexity that is independent of the size of the solution space and produces sufficient approximations of the true distribution when the solution for each pixel is most likely found in a small subset of the set of potential solutions.

PERFORMING MEDICAL TASKS BASED ON INCOMPLETE OR FAULTY DATA

A computer-implemented method and a system are for performing or supporting a medical task. An embodiment of the method includes obtaining a medical task and obtaining values for data fields of a number of available data fields. The method further includes determining whether an insufficient data field is present; and, if such a field is present, determining a relevance metric for the medical task, for the insufficient data field and/or the value thereof. Further, the method includes providing, via an estimator function, at least two different values for the insufficient data field; calculating at least two results for the medical task, which are based on the at least two different values provided; determining whether the relevance metric determined reaches or exceeds a relevance threshold value and, if this is the case, outputting an output signal based on the at least two results calculated.

Accurate ROI extraction aided by object tracking

An image data processing method includes receiving frame image data of N frames, where N>1, detecting a region of interest in one of the N frames, tracking locations of the region of interest in at least one of the N frames, and providing a merged location of the region of interest based on the locations of the region of interest in the N frames. Some embodiments include providing T of the merged locations of the region of interest for T respective groups of N frames, where T>1, providing respective statistical data for each of the T merged locations, and providing a final location of the region of interest based on the T merged locations and the statistical data for the T merged locations.

Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters.

Systems and methods for automatic scale selection in real-time imaging

A system and method for automatic scale selection in real-time image and video processing and computer vision applications. In one aspect, a non-parametric variable bandwidth mean shift technique, which is based on adaptive estimation of a normalized density gradient, is used for detecting one or more modes in the underlying data and clustering the underlying data. In another aspect, a data-driven bandwidth (or scale) selection technique is provided for the variable bandwidth mean shift method, which estimates for each data point the covariance matrix that is the most stable across a plurality of scales. The methods can be used for detecting modes and clustering data for various types of data such as image data, video data speech data, handwriting data, etc.

Pixel classification techniques

Systems, methods, and computer readable media to categorize a pixel (or other element) in an image into one of a number of different categories are described. In general, techniques are disclosed for using properties (e.g., statistics) of the regions being categorized to determine the appropriate size of window around a target pixel (element) and, when necessary, the manner in which the window may be changed if the current size is inappropriate. More particularly, adaptive window size selection techniques are disclosed for use when categorizing an image's pixels into one of two categories (e.g., black or white). Statistics of the selected region may be cascaded to determine whether the current evaluation window is acceptable and, if it is not, an appropriate factor by which to change the currently selected window's size.

EFFICIENT CALCULATION OF A ROBUST SIGNATURE OF A MEDIA UNIT

Systems, and method and computer readable media that store instructions for calculating signatures, utilizing signatures and the like.

SYSTEMS, METHODS, AND DEVICES FOR MEDICAL IMAGE ANALYSIS, DIAGNOSIS, RISK STRATIFICATION, DECISION MAKING AND/OR DISEASE TRACKING

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters.

METHOD AND SYSTEM FOR PATIENT-SPECIFIC MODELING OF BLOOD FLOW

Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three-dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

Method and system for abnormal tissue detection using z-scores in a joint histogram

Organ tissue properties of a patient are automatically compared with organ tissue properties of a healthy subject group. A population norm for the organ tissue properties is determined by: selecting at least two different tissue properties of the organ; determining for each tissue property previously selected and for each subject of said group a quantitative tissue property map; for each subject of the group, calculating a joint histogram from all the quantitative tissue property maps obtained for said subject; and determining an averaged joint histogram from all subjects of the healthy group, thus defining the population norm. A comparison is automatically performed of the averaged joint histogram with a patient joint histogram obtained for the organ tissue properties of the patient, by calculating a statistical deviation of values of a patient joint histogram relative to values of the averaged joint histogram, and mapping the statistical deviation to the patient organ.

Determining Model-Related Bias Associated with Training Data

Methods, systems, and computer program products for determining model-related bias associated with training data are provided herein. A computer-implemented method includes obtaining, via execution of a first model, class designations attributed to data points used to train the first model; identifying any of the data points associated with an inaccurate class designation and/or a low-confidence class designation; training a second model using the data points from the dataset, but excluding the identified data points; determining bias related to at least a portion of those data points used to train the second model by: modifying one or more of the data points used to train the second model; executing the first model using the modified data points; and identifying a change to one or more class designations attributed to the modified data points as compared to before the modifying; and outputting identifying information pertaining to the determined bias. 1. A computer-implemented method comprising:obtaining, in connection with execution of a first model, one or more class designations attributed to data points from a dataset used to train the first model;identifying any of the data points associated with at least one of (i) an inaccurate class designation and (ii) a class designation associated with a confidence value below a given threshold;training a second model using the data points from the dataset, but excluding the identified data points, wherein the second model is related to the first model; modifying one or more of the data points used to train the second model;', 'executing the first model using the one or more modified data points; and', 'identifying, subsequent to said executing the first model, one or more instances of bias by observing a change to one or more class designations attributed to the one or more modified data points as compared to before said modifying; and, 'determining bias related to at least a portion of those data points used to train ...

Burden Score for an Opaque Model

A method, system and computer-readable storage medium for performing a cognitive information processing operation. The cognitive information processing operation includes: receiving data from a plurality of data sources; processing the data from the plurality of data sources to provide cognitively processed insights via an augmented intelligence system, the augmented intelligence system executing on a hardware processor of an information processing system, the augmented intelligence system and the information processing system providing a cognitive computing function; performing an impartiality assessment operation via an impartiality assessment engine, the impartiality assessment operation detecting a presence of bias in an outcome of the cognitive computing function, the impartiality assessment operation generating a burden score representing the presence of bias in the outcome; and, providing the cognitively processed insights to a destination, the destination comprising a cognitive ...

Multi-task multi-modal machine learning system

Methods, systems, and apparatus, including computer programs encoded on computer storage media for training a machine learning model to perform multiple machine learning tasks from multiple machine learning domains. One system includes a machine learning model that includes multiple input modality neural networks corresponding to respective different modalities and being configured to map received data inputs of the corresponding modality to mapped data inputs from a unified representation space; an encoder neural network configured to process mapped data inputs from the unified representation space to generate respective encoder data outputs; a decoder neural network configured to process encoder data outputs to generate respective decoder data outputs from the unified representation space; and multiple output modality neural networks corresponding to respective different modalities and being configured to map decoder data outputs to data outputs of the corresponding modality.

Erkennungsvorrichtung, Erkennungsverfahren und Aufzeichnungsmedium

Spektralanalyseeinrichtung und Spektralanalyseverfahren

Eine Spektralanalyseeinrichtung 1 ist eine Einrichtung zum Analysieren eines Analyseobjekts auf Basis eines Spektrums von in dem Analyseobjekt erzeugten Licht, das ein oder zwei oder mehr einer Vielzahl von Referenzobjekten enthält, und eine Feldumwandlungseinheit 10, eine Verarbeitungseinheit 20, eine Lerneinheit 30 und eine Analyseeinheit 40 enthält. Die Feldumwandlungseinheit 10 erzeugt zwei-dimensionale Felddaten auf Basis eines Spektrums von in dem Referenzobjekt oder dem Analyseobjekt erzeugtem Licht. Die Verarbeitungseinheit 20 beinhaltet ein tiefes neuronales Netzwerk. Die Analyseeinheit 40 veranlasst die Feldumwandlungseinheit 10, die zwei-dimensionalen Felddaten auf Basis des Spektrums von in dem Analyseobjekt erzeugtem Licht zu erzeugen, gibt die zwei-dimensionalen Felddaten als Analyseobjektdaten in das tiefe neuronale Netzwerk ein und analysiert das Analyseobjekt auf Basis von aus dem tiefen neuronalen Netzwerk ausgegebenen Daten. Somit wird eine Einrichtung, die in der Lage ...

System und Verfahren zur rechnerbasierten Analyse großer Datenmengen

Für ein Computersystem zur Datenanalyse soll die Trainingszeit durch technische Vorkehrungen signifikant reduziert werden; außerdem soll der benötigte Speicherbedarf durch den Einsatz technischer Maßnahmen nennenswert sinken. Dazu wird ein elektronisches Datenverarbeitungssystem zur Analyse von Daten vorgeschlagen, mit wenigstens einem Analyse-Server und wenigstens einem Vor-Ort-Client-Rechner. Der Analyse-Server ist dazu eingerichtet und programmiert, ein selbst adaptierendes Neuronen-Netz zu implementieren, das auf eine Datenbank mit einer Vielzahl Datensätzen mit vielen Merkmalen zu trainieren ist. Der Vor-Ort-Client-Rechner ist dazu eingerichtet, ihm zugeführte Daten einer Datenvorverarbeitung und/oder einer Datenkompression zu unterziehen, bevor die Daten von dem Vor-Ort-Client-Rechner über ein elektronisches Netzwerk an den Analyse-u eingerichtet und programmiert, mit den empfangenen, vorverarbeiteten/komprimierten Daten das selbst adaptierende Neuronen-Netz zu trainieren, indem die ...

Method and system for patient-specific modeling of blood flow

C:\Users\kll\AppData\Local\Temp12713 IDAOBAFC.DOCX-22 12/2015 Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

Batch normalization layers

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer.

BATCH NORMALIZATION LAYERS

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer. Fig. 1. WO 2016/123409 PCT/US2016/015476 Neural Network Outputs Neural Network System 100 Neural Network Layer B112 Batch Normalization Layer Outputs 110 Batch Normalization Layer 108 Layer A Outputs 106 ...

Quantitative DNA-based imaging and super-resolution imaging

The present disclosure provides, ...

Saturation clustering-based method for positioning bone marrow white blood cells

A saturation clustering-based method for positioning bone marrow white blood cells: first, pre-processing a bone marrow white blood cell image to eliminate partial noise points and simultaneously smooth the image; using K-means clustering to cluster saturation channels of the bone marrow white blood cell image, and select the type of the white blood cells according to a decision tree algorithm; next, eliminating irrelevant areas in a binary image of the white blood cells by means of a morphology processing algorithm, and simultaneously filling in point holes in the white blood cells; and finally, positioning the white blood cells. The present method is simple and effective, and is suitable for a wide range of applications compared to existing threshold-based algorithms, while rendering a final result more accurate by integrating the decision tree algorithm.

A concrete durability detection method based on cloud model and D-S evidence theory

The invention discloses a concrete durability detection method based on cloud model and D-S evidence theory, which comprises the following steps: Si. Constructing a concrete durability detection index system, wherein the concrete durability detection index system comprises durability detection indexes, and durability evaluation standards and index weights corresponding to each durability detection index; S2. Based on the concrete durability detection index system, clouding the durability detection index with the cloud model, calculating the membership degree of each durability detection index corresponding to different durability grades, and normalizing the membership degree to generate evidence; S3. Integrating the evidence corresponding to each durability detection index based on the improved evidence theory DS to obtain the durability detection result of the concrete to be detected. The method can accurately detect the durability of concrete. SI- Constructing a concrete durability detection ...

A heterogeneous environment monitoring data fusion method

Offshore ship operation state identification method, device and equipment and storage medium

Information processing apparatus and method of processing information, storage medium and program

There is provided an information processing apparatus, a method of processing information, a storage medium and a program. The information processing apparatus including a statistical quantity extraction section calculating similarities between all of a group of multiple images of a first identification target and all of a group of multiple images of a second identification target and extracting a statistical quantity for similarity from the similarities and an identification section identifying the first identification target with the second identification target based on the statistical quantity for similarity. The present technology may be applied to a personal computer, for example.

METHOD FOR DETECTING DOCUMENTARY FRAUD

특징 추출 방법 및 장치

... 본 발명은 특징 추출 방법 및 장치를 개시하고, 영상 처리 기술 분야에 속한다. 상기 특징 추출 방법은, 영상을 복수 개의 셀을 포함하는 블록으로 구분하는 단계; 각 셀을 공간 영역에서 주파수 영역으로 변환하는 단계; 영상의 주파수 영역에서의 방향 그라데이션 히스토그램(HOG) 특징을 추출하는 단계를 포함한다. 주파수 영역에서 영상의 HOG 특징을 추출함으로써 HOG 특징을 추출하는 과정에서 영상의 공간 영역에 대해 직접 계산하여 얻음으로 인한 패턴 인식에서의 검출율과 정확도가 낮은 문제점를 해결하고; 주파수 영역에서 영상의 HOG 특징을 추출함으로써, 패턴 인식에서의 검출율과 정확도를 향상시키는 효과를 달성할 수 있다.

Motion control method, equipment and storage medium of the intelligent vehicle

The embodiment of the application provides a motion control method, equipment and storage medium of the intelligent vehicle, wherein the method comprises: acquiring the image to be processed; performing gesture recognition on the image to be processed to obtain gesture information of the gesture in the image to be processed; controlling the motion state of the intelligent vehicle according to the gesture information.

Systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters.

METHOD AND SYSTEM FOR IMAGE PROCESSING TO DETERMINE BLOOD FLOW

Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three-dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model.

INFORMATION RECOMMENDATION METHOD, DEVICE AND STORAGE MEDIUM

The embodiments of the present disclosure disclose an information recommendation method, device and storage medium. The method includes: determining, in a case where a user behavior is detected, an object to which the user behavior is directed as an object to be processed; determining similar objects of the object to be processed based on object similarity relationships; and recommending the similar objects; wherein the object similarity relationships are established by: acquiring labels of a plurality of sample objects; clustering the labels to obtain a plurality of label categories; for each of the sample objects, calculating similarities between a label of the sample object and the plurality of label categories to obtain a similarity set corresponding to the sample object; and establishing, according to the similarity set corresponding to each sample object, a similarity relationship between the sample object and any other one sample object of the plurality of sample objects.

IMAGE NORMALIZATION FOR FACIAL ANALYSIS

A method may include obtaining a base facial image, and obtaining a first set of base facial features within the base facial image, the first set of base facial features associated with a first facial AU to be detected in an analysis facial image. The method may also include obtaining a second set of base facial features within the base facial image, the second set of facial features associated with a second facial AU to be detected. The method may include obtaining the analysis facial image, and applying a first image normalization to the analysis facial image using the first set of base facial features to facilitate prediction of a probability of the first facial AU. The method may include applying a second image normalization to the analysis facial image using the second set of base facial features to facilitate prediction of a probability of the second facial AU. 1. A method comprising:obtaining a base facial image;obtaining a first set of base facial features within the base facial image, the first set of base facial features selected as associated with a first facial action unit (AU) to be detected in an analysis facial image;obtaining a second set of base facial features within the base facial image, at least one facial feature in the second set of base facial features being different from those in the first set of base facial features, the second set of facial features selected as associated with a second facial AU to be detected in the analysis facial image;obtaining the analysis facial image;applying a first image normalization to the analysis facial image using the first set of base facial features to facilitate prediction of a probability of the first facial AU in the analysis facial image; andapplying a second image normalization to the analysis facial image using the second set of base facial features to facilitate prediction of a probability of the second facial AU in the analysis facial image.2. The method of claim 1 , wherein applying the first ...

METHOD, APPARATUS AND COMPUTER PROGRAM PRODUCT FOR PROVIDING PATTERN DETECTION WITH UNKNOWN NOISE LEVELS

An apparatus for providing pattern detection may include a processor. The processor may be configured to iteratively test different models and corresponding scales for each of the models. The models may be employed for modeling parameters corresponding to a visually detected data. The processor may be further configured to evaluate each of the models over a plurality of iterations based on a function evaluation of each of the models, select one of the models based on the function evaluation of the selected one of the models, and utilize the selected one of the models for fitting the data. 1. A method comprising:iteratively testing, at a processor, different models and corresponding scales for each of the models, the models being employed for modeling parameters corresponding to a visually detected data, captured by an image capturing module;evaluating, by a module evaluator, each of the models over a plurality of iterations based on a function evaluation of each of the models;estimating, by a scale estimator, the scales of inlier data points for each of the models, wherein estimating the scales comprises deriving the scales based on repeated inlier data points accumulated from the different models;selecting, at the processor, one of the models based on the function evaluation of the selected one of the models; andutilizing, at the processor, the selected one of the models for fitting the data.2. The method of claim 1 , wherein utilizing the selected one of the models comprises utilizing the selected one of the models without any user input beyond provision of the data.3. The method of claim 1 , wherein evaluating each of the models comprises assigning a weighting factor to each respective model on the basis of how well each respective model fits the data and assigning a score to the model based on the weighting factor.4. The method of claim 1 , wherein selecting one of the models comprises selecting a model based on the score of the model without any provision of ...

METHOD AND SYSTEM FOR IMAGE PROCESSING TO DETERMINE BLOOD FLOW

Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three-dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model. 1184-. (canceled)185. A method for processing images to determine cardiovascular information , comprising the steps of:receiving image data including a plurality of coronary arteries originating from an aorta;processing the image data to generate three-dimensional shape models of the coronary arteries;simulating a blood flow for the generated three-dimensional shape models of the coronary arteries; anddetermining a fractional flow reserve (FFR) of the coronary arteries based on a blood flow simulation result, wherein in the step of simulating the blood flow, a computational fluid dynamics model is applied to the three-dimensional shape models of the coronary arteries, a lumped parameter model is combined with the computational fluid dynamics model, and a simplified coronary artery circulation model including coronary arteries, capillaries of the coronary arteries and coronary veins is used as the lumped parameter model.186. The method of claim 185 , wherein claim 185 , when simulating the blood flow claim 185 , when applying the computational fluid dynamics model to the three-dimensional shape models of the coronary arteries claim 185 , using an aorta blood pressure pattern as an inlet boundary condition.187. The method of claim 185 , wherein simulating the blood flow comprises determining lengths of centerlines of the three- ...

MEDICAL INFORMATION PROCESSING APPARATUS AND MEDICAL INFORMATION PROCESSING METHOD

According to one embodiment, a medical information processing apparatus has processing circuitry. The processing circuitry acquires medical data on a subject, acquires numerical data obtained by digitizing an acquisition condition of the medical data, and applies a machine learning model to input data including the numerical data and the medical data, thereby generating output data based on the medical data. 1. A medical information processing apparatus comprising a processing circuitry ,the processing circuitry being configured toacquire medical data on a subject,acquire numerical data obtained by digitizing an acquisition condition of the medical data, andapply a machine learning model to input data including the numerical data and the medical data to generate output data based on the medical data.2. The medical information processing apparatus according to claim 1 , wherein the machine learning model receives inputs of the numerical data and the medical data and learns using teacher data.3. The medical information processing apparatus according to claim 1 , whereinthe numerical data includes an element having two or more finite number of elements,the number of the elements corresponds to a number of a plurality of candidate conditions corresponding to the acquisition condition, andthe value of the element corresponds to a numerical value according to whether or not the acquisition condition is adopted for the candidate conditions and/or a value.4. The medical information processing apparatus according to claim 3 , wherein the value of the element has one of a first value indicating that the acquisition condition is adopted and a second value indicating that the acquisition condition is not adopted.5. The medical information processing apparatus according to claim 3 , whereinthe numerical data includes an element having two or more finite number of elements, andthe elements by the number of the elements include a first element corresponding to the candidate ...

METHOD FOR DETECTING AND MITIGATING BIAS AND WEAKNESS IN ARTIFICIAL INTELLIGENCE TRAINING DATA AND MODELS

Bias may be detected globally and locally by harnessing the white-box nature of the eXplainable artificial intelligence, eXplainable Neural Nets, Interpretable Neural Nets, eXplainable Transducer Transformers, eXplainable Spiking Nets, eXplainable Memory Net and eXplainable Reinforcement Learning models. Methods for detecting bias, strength, and weakness of data sets and the resulting models may be described. A method may implement global bias detection which utilizes the coefficients of the model to identify, minimize, and/or correct potential bias within a desired error tolerance. Another method makes use of local feature importance extracted from the rule-based model coefficients to locally identify bias. A third method aggregates the feature importance over the results/explanations of multiple samples. Bias may also be detected in multi-dimensional data such as images. A backmap reverse indexing mechanism may be implemented. A number of mitigation methods are also presented to eliminate bias from the affected models. 1. A method for detecting bias from an explainable model comprising at least one of a neural network , explainable artificial intelligence model , or machine learning algorithm , comprising executing on a processor the steps of:attributing values to one or more input features, wherein the input features are a plurality of features extracted from an input to the explainable model;forming an input matrix from the values of the one or more input features;extracting one or more coefficients from the explainable model,identifying one or more weights from the coefficients;forming a weight matrix, comprising the weights identified from the coefficients;multiplying the weight matrix by the input matrix to form a weighted input matrix comprising a plurality of weighted input values corresponding to the input features;identifying a highest value from the weighted input matrix and dividing each of the weighted input values by the highest value to form a ...

METHOD FOR LOCALIZATION OF BONE MARROW WHITE BLOOD CELLS BASED ON SATURATION CLUSTERING

A saturation clustering-based method for positioning bone marrow white blood cells: first, pre-processing a bone marrow white blood cell image to eliminate partial noise points and simultaneously smooth the image; using K-means clustering to cluster saturation channels of the bone marrow white blood cell image, and select the type of the white blood cells according to a decision tree algorithm; next, eliminating irrelevant areas in a binary image of the white blood cells by means of a morphology processing algorithm, and simultaneously filling in point holes in the white blood cells; and finally, positioning the white blood cells. The present method is simple and effective, and is suitable for a wide range of applications compared to existing threshold-based algorithms, while rendering a final result more accurate by integrating the decision tree algorithm. 1. A method for locating bone marrow white blood cells , comprising the steps of:(1) median filtering a bone marrow white blood cell image to remove some noise;(2) color-converting the median-filtered bone marrow white blood cell image from an RGB (red, green and blue) channel to an HSV (color, saturation, brightness) channel;(3) applying a K-means algorithm to an S-saturation channel, dividing the color-converted image into three parts, and selecting a first part or a first part and a second part to obtain a white blood cell area;(4) calculating average values (H1, H2) of an H channel in the first part and the second part obtained in step (3), and calculating the area ratio of the first part and the second part according to mean points (S1, S2) of the first part and the second part;(5) performing a statistical analysis on multiple images to identify first parts and second parts where white blood cells are included, and recording the values of H1−H2, S1−S2, and area ratios of the identified first parts and the second parts;(6) according to the recorded results in step (5), applying a decision tree algorithm to ...

AI CAPABILITY RESEARCH AND DEVELOPMENT PLATFORM AND DATA PROCESSING METHOD

Embodiments of the present disclosure provide an AI capability research and development platform and a data processing method. The AI capability research and development platform includes: a data management module, a tool management module, a process management module and a model management module, where the data management module is configured to perform data processing on received data, including at least one of the following: analyzing data type of the data, converting the data according to preset data format and storing the data; the tool management module is configured to store at least one tool, each tool being used to execute a preset processing flow; the process management module is configured to perform model training according to the tool provided by the tool management module and the data provided by the data management module; the model management module is configured to store a model obtained by the model training. 1. An artificial intelligence (AI) capability research and development platform , wherein the platform comprises:a data management module, configured to perform data processing on received data, wherein the data processing comprises at least one of the following: analyzing a data type of the data, converting the data according to a preset data format, and storing the data;a tool management module, configured to store at least one tool, each of the at least one tool being used to execute a preset processing flow;a process management module, configured to perform model training according to the tool provided by the tool management module and the data provided by the data management module; anda model management module, configured to store a model obtained by the model training.2. The platform according to claim 1 , wherein the data management module is further configured to:collect to-be-back-flow data, wherein the to-be-back-flow data is data that meets a preset back-flow condition; the to-be-back-flow data is used to provide source data for ...

MEASURING AND MONITORING SKIN FEATURE COLORS, FORM AND SIZE

Kits, diagnostic systems and methods are provided, which measure the distribution of colors of skin features by comparison to calibrated colors which are co-imaged with the skin feature. The colors on the calibration template (calibrator) are selected to represent the expected range of feature colors under various illumination and capturing conditions. The calibrator may also comprise features with different forms and size for calibrating geometric parameters of the skin features in the captured images. Measurements may be enhanced by monitoring over time changes in the distribution of colors, by measuring two and three dimensional geometrical parameters of the skin feature and by associating the data with medical diagnostic parameters. Thus, simple means for skin diagnosis and monitoring are provided which simplify and improve current dermatologic diagnostic procedures. 126-. (canceled)27. A method comprising:attaching a calibrator comprising a plurality of color areas onto a skin area next to a skin feature;capturing at least one image of the skin area, which includes both the skin feature and at least a part of the calibrator;deriving, from the at least one captured image, an image normalization function from a comparison between captured calibrator colors and the selected calibrator colors; andapplying the image normalization function to the captured skin feature to yield normalized colors of the skin feature.28. The method of claim 27 , further comprising associating at least one color parameter of the skin feature with medical diagnostic criteria claim 27 , and monitoring the at least one color parameters over time in periodically captured images.29. The method of claim 27 , further comprising capturing a plurality of images of the skin feature and deriving a three-dimensional characterization of the skin feature therefrom.30. The method of claim 27 , further comprising deriving from the at least one captured image at least one geometric parameter of the skin ...

DATA PROCESSING APPARATUS, DATA PROCESSING METHOD, AND PROGRAM

A data processing device that makes effective use of a data group containing missing data is provided. A series of learning data containing missing data is acquired, and a representative value of data and a validity ratio representing a proportion of valid data being present are calculated from the series of learning data according to a predefined unit of aggregation. Then, learning of an estimation model is performed so as to minimize an error which is based on a difference between an output resulting from inputting the representative value and the validity ratio to the estimation model, and the representative value. Also, a series of estimation data containing missing data is acquired, and a representative value of data and a validity ratio representing a proportion of valid data being present are calculated from the series of estimation data according to a predefined unit of aggregation. Then, the representative value and the validity ratio are input to the learned estimation model and a feature value is acquired or data estimation is performed for the series of estimation data. 1. A data processing device comprising:a data acquisition section, including one or more processors, configured to acquire a series of data containing missing data;a statistics calculation section, including one or more processors, configured to calculate a representative value of data and a validity ratio which represents a proportion of valid data being present from the series of data according to a predefined unit of aggregation; anda learning section, including one or more processors, configured to perform learning of an estimation model so as to minimize an error which is based on a difference between an output resulting from inputting the representative value and the validity ratio to the estimation model, and the representative value.2. The data processing device according to claim 1 , whereinthe learning section is configured to input to the estimation model an input vector made ...

METHOD AND APPARATUS FOR IMAGE RECOGNITION

The present disclosure relates to an apparatus for image recognition. The apparatus comprises a machine learning network configured to map first and second input image data to either a first or a second predefined target probability distribution, depending on whether the first and second input image data correspond to matching or non-matching images, wherein an output of the machine learning network matching the first target probability distribution is indicative of matching images and an output of the machine learning network matching the second target probability distribution is indicative of non-matching images. The present disclosure also relates to a method for training the apparatus for image recognition. 1. An apparatus for image recognition , the apparatus comprising:a machine learning network configured to map first and second input image data to either a first or a second predefined target probability distribution, depending on whether the first and second input image data correspond to matching or non-matching images,wherein an output of the machine learning network matching the first target probability distribution is indicative of matching images and an output of the machine learning network matching the second target probability distribution is indicative of non-matching images.2. The apparatus of claim 1 , wherein the first and the second target probability distribution correspond to a first and a second multivariate Gaussian distribution with distinct centers of mass.3. The apparatus of claim 1 , wherein the machine learning network comprisesa first machine learning subnetwork configured to extract respective discriminative image features from the first and second image data; anda second machine learning subnetwork configured to map the extracted first and second discriminative image features to one of the first and second predefined target probability distributions.4. The apparatus of claim 3 , wherein the first machine learning subnetwork comprises ...

SYSTEMS, METHODS, AND DEVICES FOR MEDICAL IMAGE ANALYSIS, DIAGNOSIS, RISK STRATIFICATION, DECISION MAKING AND/OR DISEASE TRACKING

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters. 1. A method for displaying computed tomography (CT) images and corresponding coronary vessel information including images rendered from the CT images , and identification and measurements of lumen , vessel walls and plaque of coronary vessels determined from the CT images by image processing , the method comprising:storing computer-executable instructions, a set of CT images of a patient's coronary vessels, and coronary vessel information associated with the set of CT images on a non-transitory computer readable medium, the coronary vessel information including information of plaque and locations of segments of the coronary vessels;generating and displaying in a user interface a first panel including a representation of coronary vessels using the coronary vessel information, the representation of coronary vessels depicting coronary vessels identified in the set of CT images;receiving a first input in the first panel indicating a selection of a coronary vessel of the representation of coronary vessels;in response to the first input, generating and displaying on the user interface a second panel illustrating at least a portion of the selected coronary ...

Anomaly Detection Method, Program, and System

A method providing an analytical technique introducing label information into an anomaly detection model. Effective utilization of label information is based on introducing the degree of similarity between samples. Assuming, for example, there is a degree of similarity between normally labeled samples and no similarity between normally labeled and abnormally labeled samples. Also each sensor value is generated by the linear sum of a latent variable and a coefficient vector specific to each sensor. However, the magnitude of observation noise is formulated to vary according to the label information for the sensor values, and set so that normal label unlabeled anomalously labeled. A graph Laplacian is created based on the degree of similarity between samples, and determines the optimal linear transformation matrix according to a gradient method. A optimal linear transformation matrix is used to calculate an anomaly score for each sensor in the test samples. 1. A computer implemented method to detect an anomaly based on measurement data , the method comprising the steps of:inputting measurement data having an anomalous or normal label and measurement data having no label as samples;determining a similarity matrix indicating the relationship between the samples based on the samples;defining a penalty based on the similarity matrix and calculating parameters in accordance with an updating equation having a term reducing the penalty; andcalculating a degree of anomaly based on the calculated parameters.2. A method according to further comprising: the step of calculating a graph Laplacian from the similarity matrix prior to the step of calculating the parameters claim 1 , the step for calculating the parameters using the calculated graph Laplacian.3. A method according to claim 2 , wherein the penalty based on the similarity is a Mahalanobis distance based on the similarity matrix or graph Laplacian.4. A method according to claim 1 , wherein the similarity matrix is a N×N ...

MACHINE-LEARNING-BASED DENOISING OF DOPPLER ULTRASOUND BLOOD FLOW AND INTRACRANIAL PRESSURE SIGNAL

An apparatus and methods for processing monitored biosignals are provided that are particularly suited for reducing noise and artifacts in continuously monitored quasi-periodic biosignals without prior knowledge of the noise distribution. The framework trains a subspace manifold with reference signals. Subsequent signals are successively projected onto the trained manifold and adjusted based on the nearest neighbors of the state of the sample being projected as well as the state of the sample at the previous time point. A denoised or modified output is obtained with inverse mapping. The reference signals may optionally be labeled during manifold training with clinical events/variables or measurable diseases/injuries from a library of relevant labels. During reconstruction, the label of the estimated state in the manifold can be obtained from the label corresponding to the estimated state. 1. An apparatus for reducing noise in continuously monitored quasi-periodic biosignals without prior knowledge of the noise distribution , comprising:(a) a computer processor; and(b) a non-transitory computer-readable memory storing instructions executable by the computer processor; (i) acquiring a plurality of reference signals;', '(ii) forming a subspace representation of the reference signals to produce a learned manifold graph;', '(iii) iteratively projecting successive signals on the learned manifold graph; and', '(iv) reconstructing the most likely shape of the successive signal., '(c) wherein said instructions, when executed by the computer processor, perform steps comprising2. The apparatus of claim 1 , wherein said instructions when executed by the computer processor further perform steps comprising:extracting individual pulses from said plurality of reference signals;distilling at least one variable from the extracted pulses;normalizing the extracted pulses; andclustering similar normalized pulses to produce an idealized reference signal.3. The apparatus of claim 1 , ...

REAL DRIFT DETECTOR ON PARTIAL LABELED DATA IN DATA STREAMS

A computerized-method for real-time detection of real concept drift in predictive machine learning models, by processing high-speed streaming data. The computerized-method includes: receiving a real-time data stream having labeled and unlabeled instances. Obtaining a window of ‘n’ instances having a portion of the ‘n’ instances as reliable labels. Computing posterior distribution of the reliable labels; and operating a Drift-Detection (DD) module. The DD module is configured to: operate a kernel density estimation on the computed posterior distribution for sensitivity control of the DD module; operate an error rate function on the estimated kernel density to yield an error value; and train an incremental estimator module, according to the kernel density estimation. When the error value is not above a preconfigured drift threshold repeating operations (i) through (iii), else when the error value is above the preconfigured drift threshold, at least one concept drift related action takes place. 2. The computerized-method of claim 1 , wherein after obtaining a window of ‘n’ instances from the data stream claim 1 , the processor is further configured to:counting the labeled instances in the ‘n’ instances;multiplying a labeling cost by the counted labeled instances to yield a total-cost;when the total-cost is not above a preconfigured labeling budget:operating a Knowledge Discovery (KD) module to obtain reliable labels of the portion of the ‘n’ instances by applying one or more machine learning models; andperforming operations (ii) through (iii).3. The computerized-method of claim 2 , before the performing of operations (ii) through (iii) claim 2 , further comprising:initiating and training of a static estimator, according to the obtained reliable labels to provide the DD module a posterior distribution.4. The computerized-method of claim 1 , wherein the reliable labels of the portion of the ‘n’ instances are provided by an end-user before the obtaining of a window of ‘n’ ...

Pixel Classification Techniques

Systems, methods, and computer readable media to categorize a pixel (or other element) in an image into one of a number of different categories are described. In general, techniques are disclosed for using properties (e.g., statistics) of the regions being categorized to determine the appropriate size of window around a target pixel (element) and, when necessary, the manner in which the window may be changed if the current size is inappropriate. More particularly, adaptive window size selection techniques are disclosed for use when categorizing an image's pixels into one of two categories (e.g., black or white). Statistics of the selected region may be cascaded to determine whether the current evaluation window is acceptable and, if it is not, an appropriate factor by which to change the currently selected window's size 1. A pixel categorization method , comprising:obtaining an image having a plurality of pixels, each pixel having a value;selecting a first pixel from the image's plurality of pixels;selecting, from the image, a first plurality of pixels associated with the first pixel;determining a first plurality of statistics based on the first plurality of pixels;identifying, from the image, a second plurality of pixels associated with the first pixel when a first statistic of the first plurality of statistics does not meet a first threshold;identifying, from the image, a third plurality of pixels associated with the first pixel when the first statistic meets the first threshold and a second statistic of the first plurality of statistics fails to meet a second threshold; andcategorizing the first pixel into one of a plurality of categories based on the first pixel's value when the first statistic meets the first threshold and the second statistic meets the second threshold.2. The method of claim 1 , wherein:a difference between a number of pixels in the first plurality of pixels and a number of pixels in the second plurality of pixels is large; anda difference ...

METHOD AND SYSTEM FOR BAD PIXEL CORRECTION IN IMAGE SENSORS

A method of detecting a bad pixel in an image sensor includes: determining, a first color of a center pixel of a Bayer patch output by the image sensor; extracting, a main patch, a first auxiliary patch, and a second auxiliary patch from the Bayer patch, having the first color, a second other color, and a third other color, respectively; generating a normalized patch of the first color from the main patch and the auxiliary patches that brings a level of the auxiliary patches to a level of the main patch; and detecting whether the center pixel of the Bayer patch is a bad pixel using the normalized patch. 1. A method of detecting a bad pixel in an image sensor , the method comprising:determining, a first color of a center pixel of a Bayer patch output by the image sensor;extracting, a main patch, a first auxiliary patch, and a second auxiliary patch from the Bayer patch, having the first color, a second other color, and a third other color, respectively;generating a normalized patch of the first color from the main patch and the auxiliary patches that brings a level of the auxiliary patches to a level of the main patch; anddetecting whether the center pixel of the Bayer patch is a bad pixel using the normalized patch.2. The method of claim 1 , wherein the detecting comprises:performing a linear interpolation on pixels of the normalized patch neighboring a center pixel of the normalized patch to generate an estimated intensity value;comparing the estimated intensity value with an actual intensity value of the center pixel of the normalized patch; anddetecting whether the center pixel of the Bayer patch is the bad pixel using a result of the comparing.3. The method of claim 2 , wherein the detecting determines the center pixel of the Bayer patch to be a bad pixel when the result indicates a difference between the actual estimated intensity value and the actual intensity value exceeds a threshold value.4. The method of claim 1 , wherein the generating of the normalized ...

Training and data synthesis and probability inference using nonlinear conditional normalizing flow model

The learning of probability distributions of data enables various applications, including but not limited to data synthesis and probability inference. A conditional non-linear normalizing flow model, and a system and method for training said model, are provided. The normalizing flow model may be trained to model unknown and complex conditional probability distributions which are at the heart of many real-life applications.

Systems, methods and devices for monitoring betting activities

System, processes and devices for monitoring betting activities using bet recognition devices and a server. Each bet recognition device has an imaging component for capturing image data for a gaming table surface. The bet recognition device receives calibration data for calibrating the bet recognition device. A server processor coupled to a data store processes the image data received from the bet recognition devices over the network to detect, for each betting area, a number of chips and a final bet value for the chips.

Image processing system and method for identifying content within image data

An image processing system for identifying content within image data. The image processing system comprises a processor that is arranged to: arrange image data in a Markov random field (MRF); and calculate state upper bound values of pairs of variables in the image data that are associated with an edge. The calculation of the state upper bound values is performed by the processor assigning the maximum state values of all of the states of a first variable of the pair of variables to the states of a second variable of the pair of variables, identifying the first and second variables from the pair of variables based on a number of states within each of the first and second variables, and determining a single state solution for identifying content in the image data based on the calculation of the state upper bound values. 1. An image processing system for identifying content within image data , the image processing system comprising a processor arranged to:arrange image data in a Markov random field (MRF);calculate state upper bound values of pairs of variables in the image data that are associated with an edge,wherein the processor is further arranged to calculate the state upper bound values by assigning maximum state values of all of the states of a first variable of the pair of variables to the states of a second variable of the pair of variables, and identifying the first and second variables from the pair of variables based on a number of states within each of the first and second variables, anddetermine a single state solution for identifying content in the image data based on the calculation of the state upper bound values.2. The image processing system of claim 1 , wherein the processor is further arranged to:determine whether one variable in the pair of variables has more states than the other variable in the pair of variables, andidentify the first and second variables from each pair of variables based on the determination.3. The image processing system of ...

GROUND EXTRACTION METHOD FOR 3D POINT CLOUDS OF OUTDOOR SCENES BASED ON GAUSSIAN PROCESS REGRESSION

A ground extraction method for 3D point clouds of outdoor scenes based on Gaussian process regression, including: (1) obtaining the 3D point cloud of an outdoor scene, (2) building the neighborhood of the 3D point cloud, (3) calculating the covariance matrices and normal vectors of the 3D point cloud, (4) classifying the 3D point cloud according to its neighborhood shape, (5) extracting the initial ground G, (6) segmenting the initial ground, (7) 2D Gaussian process regression, (8) finding the neighborhood Nof each ground fragment LG, and (9) extracting the final ground G. 1step 1: obtaining a 3D point cloud of an outdoor scene by using a laser rangefinder, wherein the 3D point cloud of the outdoor scene is a set of discrete points;{'sub': i', 'i', 'i', 'i', 'n', 'i', 'i', 'n', 'i, 'step 2: building a neighborhood of the 3D point cloud, comprising: constructing a structure tree of the 3D point cloud by using a KD-tree algorithm, dividing the 3D point cloud into different spatial regions according to coordinates of the discrete points in the 3D point cloud, using spatial address information to search neighboring points in the process of the neighborhood construction, and building the neighborhood N={p=(x,y,z)|1≤i≤n} of a given point p=(x,y,z) in the 3D point cloud, wherein pis a neighboring point of the given point p=(x,y,z), i is the serial number of the neighboring points p, and nis the number of the neighboring points p;'}{'sub': i', 'i', 'i', 'i', 'n', '1', '2', '3', '1', '2', '3, 'step 3: calculating covariance matrices and normal vectors of the 3D point cloud, wherein for the given point p=(x,y,z) in the 3D point cloud, a covariance matrix M is constructed by using its neighborhood N={p=(x,y,z)|1≤i≤n}, and eigenvalues λ,λ,λand eigenvectors v,v,vof the covariance matrix M and a normal vector n of the given point p are computed; and step 3 comprises the following substeps{'sub': i', 'i', 'i', 'i', 'n, '(a) building the neighborhood N={p=(x,y,z)|1≤i≤n} of the ...

COMBINATORIAL BAYESIAN OPTIMIZATION USING A GRAPH CARTESIAN PRODUCT

Certain aspects provide a method for determining a solution to a combinatorial optimization problem, including: determining a plurality of subgraphs, wherein each subgraph of the plurality of subgraphs corresponds to a combinatorial variable of the plurality of combinatorial variables; determining a combinatorial graph based on the plurality of subgraphs; determining evaluation data comprising a set of vertices in the combinatorial graph and evaluations on the set of vertices; fitting a Gaussian process to the evaluation data; determining an acquisition function for vertices in the combinatorial graph using a predictive mean and a predictive variance from the fitted Gaussian process; optimizing the acquisition function on the combinatorial graph to determine a next vertex to evaluate; evaluating the next vertex; updating the evaluation data with a tuple of the next vertex and its evaluation; and determining a solution to the problem, wherein the solution comprises a vertex of the combinatorial graph. 1. A method for optimizing a plurality of combinatorial variables , comprising:receiving a plurality of combinatorial variables as input to an optimization process;determining a plurality of subgraphs, wherein each subgraph of the plurality of subgraphs corresponds to a combinatorial variable of the plurality of combinatorial variables;determining a combinatorial graph based on the plurality of subgraphs, wherein the combinatorial graph comprises a plurality of vertices;determining evaluation data comprising a set of vertices from the plurality of vertices in the combinatorial graph and evaluations on the set of vertices;fitting a Gaussian process to the evaluation data, wherein the fitted Gaussian process comprises a predictive mean and a predictive variance;determining an acquisition function for vertices in the combinatorial graph using the predictive mean and the predictive variance from the fitted Gaussian process;adjusting the acquisition function on the ...

INTENT-BASED TELEMETRY COLLECTION SERVICE

A telemetry service can receive telemetry collection requirements that are expressed as an “intent” that defines how telemetry is to be collected. A telemetry intent compiler can receive the telemetry intent and translate the high level intent into abstract telemetry configuration parameters that provide a generic description of desired telemetry data. The telemetry service can determine, from the telemetry intent, a set of devices from which to collect telemetry data. For each device, the telemetry service can determine capabilities of the device with respect to telemetry data collection. The capabilities may include a telemetry protocol supported by the device. The telemetry service can create a protocol specific device configuration based on the abstract telemetry configuration parameters and the telemetry protocol supported by the device. Devices in a network system that support a particular telemetry protocol can be allocated to instances of a telemetry collector that supports the telemetry protocol. 1. A method for providing a telemetry service in a virtualized computing infrastructure , the method comprising:receiving, by one or more processors, data representing telemetry data collection intent;translating, by the one or more processors, the data representing the telemetry data collection intent into one or more abstract telemetry configuration parameters;determining, by the one or more processors and from the data representing the telemetry collection intent, a set of devices of the virtualized computing infrastructure;obtaining, by the one or more processors, device capability information for the set of devices, the device capability information including, for each device of the set of devices, a supported telemetry protocol; andfor each device of the set of devices, configuring the device based on the abstract telemetry configuration parameters and allocating the device to an instance of a plurality of telemetry collectors based on the supported telemetry ...

METHODS AND APPARATUS TO GENERATE AUDIENCE METRICS USING MATRIX ANALYSIS

An example apparatus includes audience metrics collecting circuitry to access first audience metrics from a server, and access second audience metrics from the server, matrix building circuitry to build a matrix of the first audience metrics and the second audience metrics, missing values of the matrix corresponding to the second audience metrics, data transforming circuitry to transform the first audience metrics and the second audience metrics in the matrix, missing value calculating circuitry to determine imputed transformed values of the missing values using a recommender system, and the data transforming circuitry to recover imputed values of the missing values based on the imputed transformed values. 110.-. (canceled)11. An apparatus comprising:at least one memory;instructions in the apparatus; and{'claim-text': ['access first audience metrics from a server;', 'access second audience metrics from the server;', 'build a matrix of the first audience metrics and the second audience metrics, missing values of the matrix corresponding to the second audience metrics;', 'transform the first audience metrics and the second audience metrics in the matrix;', 'determine imputed transformed values of the missing values using a recommender system; and', 'recover imputed values of the missing values based on the imputed transformed values.'], '#text': 'processor circuitry to execute the instructions to:'}12. The apparatus of claim 11 , wherein the first audience metrics include a first audience size claim 11 , a first impression count claim 11 , and first duration data claim 11 , and the second audience metrics include a second audience size claim 11 , a second impression count claim 11 , and second duration data.13. The apparatus of claim 11 , wherein a first portion of the first audience metrics corresponds to panelist audience members for whom first demographic information is known and a second portion of the first audience metrics corresponds to census audience members ...

MANAGING MISSING VALUES IN DATASETS FOR MACHINE LEARNING MODELS

Computer-implemented machines, systems and methods for managing missing values in a dataset for a machine learning model. The method may comprise importing a dataset with missing values; computing data statistics and identifying the missing values; verifying the missing values; updating the missing values; imputing missing values; encoding reasons for why values are missing; combining imputed missing values and the encoded reasons; and recommending models and hyperparameters to handle special or missing values. 1. A computer-implemented method for managing missing values in a dataset for a machine learning model , wherein the machine learning model is trained , during a training phase , to learn patterns to correctly classify input data associated with risk analysis or risk predication , the method comprising:importing a dataset with data point values, the dataset being usable for training the machine learning model, one or more of the data point values being associated with one or more features of the machine learning model;in response to determining that at a first data point value associated with a first feature is missing, imputing a value for the first data point value based on an imputation method selected from a group of imputation options; andupdating the first data point value in accordance with results obtained from applying a selected imputations method.2. The method of claim 1 , wherein data statistics are computed to determine that the first data point value is missing.3. The method of claim 2 , wherein in response to interaction with a domain expert claim 2 , it is verified that the first data point value is missing.4. The method of claim 1 , wherein an encoding process is utilized to provide reasons for the first data point missing from the dataset.5. The method of claim 4 , wherein the provided reasons and the imputed value for the first data point and a corresponding reason for the first data point missing from the dataset are combined.6. The method ...

DUPLICATE/NEAR DUPLICATE DETECTION AND IMAGE REGISTRATION

Embodiments are disclosed for detecting duplicate and near duplicate images. An exemplary method includes receiving an original image, preparing the image for fingerprinting, and calculating an image fingerprint, the fingerprint expressed as a sequence of numbers. The method further includes comparing the image fingerprint thus obtained with a set of previously stored fingerprints obtained from a set of previously stored images, and determining if the original image is either a duplicate or a near duplicate of an image in the set if the dissimilarity between the two fingerprints is less than a defined threshold T. Once a duplicate or near duplicate is detected, various defined actions may be taken, including culling the less desirable image or referring the redundancy to a user. 1. A method , comprising:generating, using at least one processor, a cell array for an image, the cell array comprising a grid of cells corresponding to regions of the image, the grid of cells comprising numeric values representative of pixel intensity values of the corresponding regions of the image;generating an image fingerprint for the image, the image fingerprint comprising a sequence of the numeric values of the cell array;comparing the image fingerprint for the image to a plurality of previously generated image fingerprints for a plurality of images to identify, from the plurality of images, a similar image to the image; andin response to identifying the similar image to the image, taking a defined action with respect to the similar image.2. The method of claim 1 , further comprising claim 1 , prior to generating the cell array for the image claim 1 , normalizing the image to a standard orientation.3. The method of claim 2 , wherein normalizing the image to the standard orientation comprises:generating a normalization cell array comprising numeric values representative of pixel intensity values of the image; androtating the image to the standard orientation based on the numeric values ...

Variance Characterization Based on Feature Contribution

Systems, methods, and computer readable media are disclosed for generating, modifying, and using machine learning models to predict and evaluate variances between data sets. Methods disclosed herein may include identifying features that characterize members of a data set, generating a machine learning model using identified features, using the machine learning model and the group to assign feature attributions to the features, and predicting the impact of those features on behaviors of the first data set and a second data set.

CONVERSATION HISTORY WITHIN CONVERSATIONAL MACHINE READING COMPREHENSION

Aspects described herein include a method of conversational machine reading comprehension, as well as an associated system and computer program product. The method comprises receiving a plurality of questions relating to a context, and generating a sequence of context graphs. Each of the context graphs includes encoded representations of: (i) the context, (ii) a respective question of the plurality of questions, and (iii) a respective conversation history reflecting: (a) one or more previous questions relative to the respective question, and (b) one or more previous answers to the one or more previous questions. The method further comprises identifying, using at least one graph neural network, one or more temporal dependencies between adjacent context graphs of the sequence. The method further comprises predicting, based at least on the one or more temporal dependencies, an answer for a first question of the plurality of questions. 1. A method of conversational machine reading comprehension , the method comprising:receiving a plurality of questions relating to a context;generating a sequence of context graphs, wherein each of the context graphs includes encoded representations of: (i) the context, (ii) a respective question of the plurality of questions, and (iii) a respective conversation history reflecting: (a) one or more previous questions relative to the respective question, and (b) one or more previous answers to the one or more previous questions;identifying, using at least one graph neural network, one or more temporal dependencies between adjacent context graphs of the sequence; andpredicting, based at least on the one or more temporal dependencies, an answer for a first question of the plurality of questions.2. The method of claim 1 , wherein generating the sequence of context graphs comprises:encoding each question of the plurality of questions, wherein encoding each question comprises encoding each word of the question with one or more of: one or more ...

APPARATUS AND METHOD FOR SINGLE LOOK MAIN LOBE AND SIDELOBE DISCRIMINATION IN SPECTRAL DOMAIN IMAGES

A system performs operations including receiving multi-dimensional single-look data from a sensor, applying multi-dimensional complex weighting functions including apodizations from among a general class of such functions to the complex data, so as to induce nonlinear variations in the amplitude and phase of the multi-dimensional spectral image responses, forming a number of features per voxel across a number of multi-dimensional spectral image responses, and using a multi-dimensional non-parametric classifier to form features to discriminate main lobe from sidelobe imaged voxels with the weighting function applied to received data. The operations include identifying each voxel by processing a set of transforms from the multi-dimensional complex weighting functions and outputting a multi-dimensional main lobe binary image, representing main lobe versus sidelobe locations. 1. A system comprising: receiving multi-dimensional single-look data from a sensor;', 'applying a number of multi-dimensional complex weighting functions including apodizations from among a general class of such functions to the complex data, so as to induce nonlinear variations in the amplitude and phase of the multi-dimensional spectral image responses;', 'forming a number of features per voxel across a number of multi-dimensional spectral image responses;', 'using a multi-dimensional non-parametric classifier to form statistics from features to discriminate main lobe from sidelobe imaged voxels with the weighting function applied to received data;', 'identifying each voxel by thresholding classifier statistics; and', 'outputting a multi-dimensional main lobe binary image, representing main lobe versus sidelobe locations., 'a computing device comprising a memory configured to store instructions and a processor to execute the instructions to perform operations comprising2. The system of claim 1 , in which the multi-dimensional non-parametric classifier operations include:processing a set of ...

PIXEL-BASED LOAD BALANCING

Disclosed are methods, systems, and apparatus for load-balancing image-processing jobs based on the number of pixels in the images and/or the nature of the processing that is requested on those pixels. For example, a set of machines may run software to provide various types of image processing services, such as format conversion, recompression, resizing, cropping, among others. These are referred to as image servers. In accordance with the teachings hereof, the load on each image server can be calculated based on the number of pixels in the images that are waiting to be processed in the image server's processing queue, adjusted by the type of processing that is requested on each image. The adjustment typically reflects and adjusts for the relative time needed to perform various types of processing. Load scores can be further adjusted based on the processing capabilities of each image server, in some embodiments. 121.-. (canceled)22. A method of load-balancing image processing requests across a plurality of image servers in a distributed computing system that provides a content delivery network (CDN) , the method comprising:receiving a request to process a first image, the request comprising an identification of the first image and a directive to process the first image in accordance with a first image processing service;determining a value representative of a number of pixels in the first image, wherein said determination comprises at least one of: (a) counting the number of pixels in the first image and (b) reading pixel information from an HTTP header of the request;determining a pixel load for each of the plurality of image servers, each pixel load based at least in part on a value representative of a number of pixels in queue for a respective image server in the plurality of image servers;determining a pixel drain rate for each of the plurality of image servers, each pixel drain rate based at least in part on a number of pixels processed during a time period by ...

GENERATING A HIERARCHY OF VISUAL PATTERN CLASSES

A hierarchy machine may be configured as a clustering machine that utilizes local feature embedding to organize visual patterns into nodes that each represent one or more visual patterns. These nodes may be arranged as a hierarchy in which a node may have a parent-child relationship with one or more other nodes. The hierarchy machine may implement a node splitting and tree-learning algorithm that includes hard-splitting of nodes and soft-assignment of nodes to perform error-bounded splitting of nodes into clusters. This may enable the hierarchy machine, which may form all or part of a visual pattern recognition system, to perform large-scale visual pattern recognition, such as font recognition or facial recognition, based on a learned error-bounded tree of visual patterns. 1. A method comprising:classifying a reference set of visual patterns that belong to a parent class into mutually exclusive child classes that include first and second child classes, a visual pattern from the reference set being classified into the first child class instead of the second child class;modifying a weight vector that corresponds to the parent class, the modified weight vector altering a first probability that the visual pattern belongs to the first child class and a second probability that the visual pattern belongs to the second child class;based on the altered first and second probabilities, removing mutual exclusivity from the first and second child classes by adding the visual pattern to the second child class; andusing a processor, generating a hierarchy of classes of visual patterns, the hierarchy including the parent class and the mutually nonexclusive first and second child classes that each include the visual pattern.2. The method of claim 1 , wherein the classifying of the reference set includes:computing an affinity matrix that quantifies degrees to which the visual patterns that belong to the parent class are similar to each other; andgrouping the visual patterns into the ...

DEVICE, SYSTEM AND METHOD FOR SKIN DETECTION

The present invention relates to a device, system and method for skin detection. To enable a reliable, accurate and fast detection the proposed device comprises an input interface () for obtaining image data of a scene, said image data comprising a time sequence of image frames, an extraction unit () for extracting a photoplethysmography (PPG) signal from a region of interest of said image data, a transformation unit () for transforming said PPG signal into a spectral signal, a sorting unit () for sorting said spectral signal to obtain a sorted spectral signal representing a descriptor, and a classifier () for classifying said region of interest as skin region of a living being or as non-skin region based on the descriptor. 1. A device for skin detection comprising:an input interface for obtaining image data of a scene, said image data comprising a time sequence of image frames,an extraction unit for extracting a photoplethysmography, PPG, signal from a region of interest of said image data,a transformation unit for transforming said PPG signal into a spectral signal,a sorting unit for sorting said spectral signal to obtain a sorted spectral signal representing a descriptor, anda classifier for classifying said region of interest as skin region of a living being or as non-skin region based on the descriptor.2. The device as claimed in claim 1 ,wherein said transformation unit is configured to transform said PPG signal into a spectral signal without phase information, in particular into a power spectrum or an absolute spectrum.3. The device as claimed in claim 1 ,wherein said sorting unit is configured to divide said spectral signal into two or more spectral sub-signals and to separately sort said sub-signals to obtain sorted spectral sub-signals representing the descriptor.4. The device as claimed in claim 1 ,wherein said sorting unit is configured to divide said spectral signal into an in-band sub-signal covering a first frequency range of said spectral signal and ...

Unsupervised Deep Representation Learning for Fine-grained Body Part Recognition

A method and apparatus for deep learning based fine-grained body part recognition in medical imaging data is disclosed. A paired convolutional neural network (P-CNN) for slice ordering is trained based on unlabeled training medical image volumes. A convolutional neural network (CNN) for fine-grained body part recognition is trained by fine-tuning learned weights of the trained P-CNN for slice ordering. The CNN for fine-grained body part recognition is trained to calculate, for an input transversal slice of a medical imaging volume, a normalized height score indicating a normalized height of the input transversal slice in the human body. 1. A method for deep learning based fine-grained body part recognition in medical imaging data , comprising:training a paired convolutional neural network (P-CNN) for slice ordering based on unlabeled training medical image volumes; andtraining a convolutional neural network (CNN) for fine-grained body part recognition by fine-tuning learned weights of the trained P-CNN for slice ordering.2. The method of claim 1 , wherein training a paired convolutional neural network (P-CNN) for slice ordering based on unlabeled training medical image volumes comprises:randomly sampling transversal slice pairs from the unlabeled medical image training volumes, wherein each transversal slice pair is randomly sampled from the same training volume; andtraining the P-CNN to predict a relative order of a pair of transversal slices of a medical imaging volume based on the randomly sampled transversal slice pairs, wherein the P-CNN includes two identical sub-networks for a first plurality of layers, each to extract feature from a respective slice of the pair of transversal slices, and global final layers to fuse outputs of the sub-networks and calculate a binary classification result regarding the relative order of the pair of transversal slices.3. The method of claim 2 , wherein the CNN for fine-grained body part recognition includes a first plurality of ...

Image classification using batch normalization layers

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing images or features of images using an image classification system that includes a batch normalization layer. One of the systems includes a convolutional neural network configured to receive an input comprising an image or image features of the image and to generate a network output that includes respective scores for each object category in a set of object categories, the score for each object category representing a likelihood that that the image contains an image of an object belonging to the category, and the convolutional neural network comprising: a plurality of neural network layers, the plurality of neural network layers comprising a first convolutional neural network layer and a second neural network layer; and a batch normalization layer between the first convolutional neural network layer and the second neural network layer.

Co-Expression Signatures Method for Quantification of Physiological and Structural Data

Described here are systems and methods for generating and analyzing co-expression signature data from scalar or multi-dimensional data fields contained in or otherwise derived from imaging data acquired with a medical imaging system. A similarity metric, such as an angular similarity metric, is computed between the data field components contained in pairs of voxels in the data field data. The data fields can be scalar fields, vector fields, tensor fields, or other higher-dimensional data fields. A probability distribution of these similarity metrics can be generated and used as co-expression signature data that indicate pairwise disparities in the data field data. 1. A method for generating co-expression signature data from data field data obtained from imaging data acquired with an imaging system , the method comprising:(a) accessing data field data comprising a plurality of voxels, wherein the data field data were obtained from imaging data acquired from a subject using an imaging system and each voxel in the data field data comprises data field component; computing a similarity metric for each of a plurality of voxel pairs in the data field data, wherein the similarity metric indicates one of a similarity or disparity between the data field component in each voxel pair; and', 'computing the co-expression signature data as a distribution of the similarity metrics;, '(b) generating co-expression signature data from the data field data bywherein the co-expression signature data encode a distribution of pairwise similarities or disparities in the data field data that are representative of at least one of physiological changes or structural changes in the subject.2. The method as recited in claim 1 , further comprising generating signature dissimilarity data by comparing the co-expression signature data with baseline data.3. The method as recited in claim 2 , wherein the co-expression signature data are compared with the baseline data by computing a distance metric ...

IMAGE PROCESSING DEVICE, IMAGE FORMING APPARATUS, IMAGE PROCESSING METHOD, AND NON-TRANSITORY COMPUTER-READABLE STORAGE MEDIUM

An image processing device includes: an image classifying section which, through a convolutional neural network, classifies each pixel of input image data as expressing or not expressing a handwritten image to calculate a classification probability of each pixel, the classification probability being a probability that the handwritten image is expressed; a threshold setting section which sets a first threshold when removal processing to remove the handwritten image is performed and a second threshold which is smaller than the first threshold when emphasis processing to emphasize the handwritten image is performed; and an image processor which adjusts a gradation value of pixels with a classification probability no smaller than the first threshold to remove the handwritten image when the removal processing is performed and adjusts the gradation value of pixels with a classification probability no smaller than the second threshold to emphasize the handwritten image when the emphasis processing is performed. 1. An image processing device comprising:an image classifying section configured to, through a convolutional neural network, classify each pixel of input image data as expressing or not expressing a handwritten image to calculate a classification probability of each pixel, the classification probability being a probability that the handwritten image is expressed;a threshold setting section configured to set a first threshold when removal processing is performed and a second threshold when emphasis processing is performed, the removal processing being image processing to remove the handwritten image, the emphasis processing being image processing to emphasize the handwritten image, the second threshold being smaller than the first threshold; andan image processor configured to adjust a gradation value of pixels for which the classification probability is no smaller than the first threshold to remove the handwritten image when the removal processing is performed and ...

REAL-TIME MINIMAL VECTOR LABELING SCHEME FOR SUPERVISED MACHINE LEARNING

Systems and methods are described for training a machine learning model using intelligently selected multiclass vectors. According to an embodiment, a set of un-labeled feature vectors are received. The set of feature vectors are grouped into clusters within a vector space having fewer dimensions than the first set of feature vectors by applying a homomorphic dimensionality reduction algorithm to the set of feature vectors and performing centroid-based clustering. An optimal set of clusters among the clusters is identified by performing a convex optimization process on the clusters. Vector labeling is minimized by selecting ground truth representative vectors including a representative vector from each cluster of the optimal set of clusters. A set of labeled feature vectors is created based on labels received from an oracle for each of the representative vectors. A machine-learning model is trained for multiclass classification based on the set of labeled feature vectors. 1. A method comprising:receiving, by a processing resource of a computing system, a first set of feature vectors, wherein the first set of feature vectors are un-labeled;grouping, by the processing resource, the first set of feature vectors into a plurality of clusters within a vector space having fewer dimensions than the first set of feature vectors by applying a homomorphic dimensionality reduction algorithm to the first set of feature vectors and performing centroid-based clustering;identifying, by the processing resource, an optimal set of clusters among the plurality of clusters by performing a convex optimization process on the plurality of clusters;minimizing, by the processing resource, vector labeling by selecting a plurality of ground truth representative vectors including a representative vector from each cluster of the optimal set of clusters;creating, by the processing resource, a set of labeled feature vectors based on labels received from an oracle for each of the plurality of ...

INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, AND PROGRAM

An apparatus includes a distribution generator configured to determine a combination of each of first regions in a first image and a corresponding one of second regions in a second image based on feature amounts of the first regions and feature amounts of the second regions, and generate first distribution information indicating distribution of the difference between a first position of a first region in the first image and a second position of a second region in the second image; a similarity generator configured to generate second distribution information indicating distribution of the difference between the first position and the second position, calculate a first similarity between the first image and the second image, and calculate a second similarity between the first image and the second image; and a processor configured to generate a comparison result by comparing the first image and the second image. 1. An information processing device comprising:a distribution generator configured to determine a combination of each of a plurality of regions in a first image and a corresponding one of a plurality of regions in a second image based on feature amounts, represented by a string of bits, of the regions in the first image and feature amounts of the regions in the second image, and generate first distribution information indicating distribution of the difference between a position of a first region in the first image and a position of a second region in the second image that are included in each of the combinations;a similarity generator configured to generate second distribution information indicating distribution of the difference between the position of the first region and the position of the second region by applying a spatial filter to the first distribution information, calculate a first similarity between the first image and the second image based on the first distribution information, and calculate a second similarity between the first image and the ...

OBJECT DETECTION AND LEARNING METHOD AND APPARATUS

A processor-implemented object detection method is provided. The method receives an input image, generates a latent variable that indicates a feature distribution of the input image, and detects an object in the input image based on the generated latent variable. 1. A processor-implemented object detection method comprising:receiving an input image;generating a latent variable that indicates a feature distribution of the input image; anddetecting an object in the input image based on the generated latent variable.2. The method of claim 1 , wherein the generating of the latent variable comprises:extracting a feature of the input image with a neural network; andacquiring the latent variable with the neural network,wherein the neural network encodes the input image and determines a feature distribution corresponding to the feature of the input image.3. The method of claim 2 , wherein the neural network is trained based on at least one of a real image and a synthetic image that is filtered based on a score of a transformed synthetic image that is acquired by transforming the synthetic image to a fake image of the real image.4. The method of claim 1 , wherein the latent variable comprises a feature used to translate a domain of the input image and a feature used to detect the object.5. The method of claim 4 , wherein the feature used to translate the domain of the input image comprises a feature that is shared between a feature used to translate a first domain of a synthetic image to a second domain of a real image claim 4 , and a feature used to translate the second domain to the first domain.6. The method of claim 1 , wherein the latent variable comprises a multi-dimensional mean vector and a multi-dimensional distribution vector.7. The method of claim 1 , wherein the detecting of the object in the input image comprises:acquiring information indicating a location of the object in the input image; andacquiring information that classifies the object.8. A learning method ...

METHOD AND SYSTEM FOR IMAGE PROCESSING TO DETERMINE BLOOD FLOW

Embodiments include a system for determining cardiovascular information for a patient. The system may include at least one computer system configured to receive patient-specific data regarding a geometry of the patient's heart, and create a three-dimensional model representing at least a portion of the patient's heart based on the patient-specific data. The at least one computer system may be further configured to create a physics-based model relating to a blood flow characteristic of the patient's heart and determine a fractional flow reserve within the patient's heart based on the three-dimensional model and the physics-based model. 1184-. (canceled)185. A medical image diagnostic apparatus for processing medical images , comprising:processing circuitry configured to acquire data of a plurality of Fractional Flow Reserve (FFR) distribution maps comprising at least two time phases regarding a coronary artery; anda display configured to display a representation of the at least one FFR distribution map in phase to the at least one FFR distribution map, wherein the processing circuitry restricts display objects displayed by the display for the representation of the at least one FFR distribution map based on the at least one FFR distribution map.186. The medical image diagnostic apparatus according to claim 185 , wherein the processing circuitry acquires data of a plurality of morphological images comprising the at least two time phases claim 185 , and converts the at least one FFR distribution map into a at least one corresponding color map claim 185 , respectively claim 185 , and the display displays a plurality of superimposed images claim 185 , as the representation of the at least one FFR distribution map claim 185 , obtained by superposing the at least one color map and the plurality of morphological images respectively corresponding in phase.187. The medical image diagnostic apparatus according to claim 186 , wherein the processing circuitry restricts display ...

SITUATION AWARENESS AND DYNAMIC ENSEMBLE FORECASTING OF ABNORMAL BEHAVIOR IN CYBER-PHYSICAL SYSTEM

A plurality of monitoring nodes may each generate a time-series of current monitoring node values representing current operation of a cyber-physical system. A feature-based forecasting framework may receive the time-series of and generate a set of current feature vectors using feature discovery techniques. The feature behavior for each monitoring node may be characterized in the form of decision boundaries that separate normal and abnormal space based on operating data of the system. A set of ensemble state-space models may be constructed to represent feature evolution in the time-domain, wherein the forecasted outputs from the set of ensemble state-space models comprise anticipated time evolution of features. The framework may then obtain an overall features forecast through dynamic ensemble averaging and compare the overall features forecast to a threshold to generate an estimate associated with at least one feature vector crossing an associated decision boundary. 1. A system to protect a cyber-physical system , comprising:a plurality of monitoring nodes each generating a time-series of current monitoring node values that represent a current operation of the cyber-physical system; and receive the time-series of current monitoring node values and generate a set of current feature vectors using feature discovery techniques,', 'characterize the feature behavior for each monitoring node in the form of decision boundaries that separate normal and abnormal space based on operating data of the cyber-physical system,', 'construct a set of ensemble state-space models to represent feature evolution in the time-domain, wherein the forecasted outputs from the set of ensemble state-space models comprise anticipated time evolution of features,', 'obtain an overall features forecast through dynamic ensemble averaging,', 'compare the overall features forecast to a threshold to generate an estimate associated with at least one feature vector crossing an associated decision ...

Information processing device, information processing method, and storage medium

An information processing device includes: a statistics unit that calculates an input data amount within a predetermined period for stream data which is divided into a plurality of divided data and on which distributed processing is performed; and a determination unit that determines a divided duration of the stream data based on the input data amount so that the number of times of transfer of the divided data between a plurality of nodes when the distributed processing is performed by the plurality of nodes satisfies a predetermined condition.

IMAGE RECONSTRUCTION SYSTEM AND METHOD IN MAGNETIC RESONANCE IMAGING

A method and system for image reconstruction are provided. Multiple coil images may be obtained. A first reconstructed image based on the multiple coil images may be reconstructed based on a first reconstruction algorithm. A second reconstructed image based on the multiple coil images may be reconstructed based on a second reconstruction algorithm. Correction information about the first reconstructed image may be generated based on the first reconstructed image and the second reconstructed image. A third reconstructed image may be generated based on the first reconstructed image and the correction information about the first reconstructed image. 1. A method for reconstructing a corrected image implemented on a magnetic resonance imaging (MRI) system including an MRI device and a computing device , the MRI device including multiple radio frequency (RF) receiver coils , the computing device including a processor , the method comprising:receiving, by the multiple RF receiver coils, MR signals of an object;generating, by the processor, multiple coil images of the object based on the MR signals, the multiple coil images including a first set of coil images and a second set of coil images;reconstructing, by the processor, a first reconstructed image based on the first set of coil images according to a first reconstruction algorithm;reconstructing, by the processor, a second reconstructed image based on the second set of coil images according to a second reconstruction algorithm, the second reconstruction algorithm being different with the first reconstruction algorithm;generating, by the processor, correction information about the first reconstructed image by dividing the first reconstructed image by the second reconstructed image; andgenerating, by the processor, the corrected image with reduced inhomogeneity intensity based on the first reconstructed image and the correction information about the first reconstructed image.2. The method of claim 1 , the generating ...

FOOD PREPARATION ENTITY

The invention relates to a food preparation entity comprising a cavity () for receiving food to be prepared and an image recognition system () for gathering optical information of the food to be prepared, wherein the food preparation entity () is further adapted to store, gather and/or receive meta-information and select one or more food types out of a list of food types based on said meta-information and said captured optical information. 1. Food preparation entity comprising a cavity for receiving food to be prepared and an image recognition system for capturing optical information of the food to be prepared , wherein the food preparation entity is further adapted to store , gather and/or receive meta-information and select one or more food types out of a list of food types based on said meta-information and said captured optical information.2. Food preparation entity according to claim 1 , comprising a processing entity adapted to perform a food preselection based on the captured optical information in order to determine a subset of possible food types which may be received within the cavity claim 1 , wherein the food preparation entity is further adapted to select one or more food types out of the subset of possible food types based on said meta information.3. Food preparation entity according to claim 1 , adapted to store claim 1 , gather and/or receive geographical information and the food preparation entity is further adapted to select one or more food types out of the subset of possible food types based on said geographical information.4. Food preparation entity according to claim 2 , adapted to associate each food included in the subset of possible food types with a weighting factor claim 2 , said weighting factor depending on the geographical information and indicating the a frequency of consumption of said food in a geographical region characterized by said geographical information.5. Food preparation entity according to claim 1 , wherein said meta- ...

DATA ANALYSIS DEVICE, METHOD, AND PROGRAM

There are provided a data analysis device, a method, and a program that are capable of improving the accuracy of predicting an output variable for an unknown input variable by making it possible to use input/output data in which the value of the output variable is given as an interval value. A data analysis device A includes: a data processing unit that performs a process of acquiring data represented by a set of a plurality of first input/output data in which a value of an output variable is given and a plurality of second input/output data in which a value of an output variable is gives as an interval value representing a range; and a prediction unit that, based on an input variable for which a value of an output variable is unknown and the data, predicts a value of an output variable for the unknown input variable using a Gaussian process. 1. A data analysis device comprising:a data processing unit that performs a process of acquiring data represented by a set of a plurality of first input/output data in which a value of an output variable is given and a plurality of second input/output data in which a value of an output variable is given as an interval value representing a range; anda prediction unit that, based on an input variable for which a value of an output variable is unknown and the data, predicts a value of an output variable for the unknown input variable using a Gaussian process.2. The data analysis device according to claim 1 , further comprisinga latent variable estimation unit that estimates a latent variable representing an estimate of a true value of an output variable given as the interval value for each of the second input/output data,the latent variable estimation unit generating a random number as the latent variable according to a truncated normal distribution of a generation probability of a latent variable conditioned by the interval value, the truncated normal distribution being represented using a kernel function that represents ...

QUANTUM ERROR MITIGATION USING HARDWARE-FRIENDLY PROBABILISTIC ERROR CORRECTION

Techniques facilitating error mitigation for quantum computing devices. In one example, a system can comprise a process that executes computer executable components stored in memory. The computer executable components comprise: an approximation component; a budget component; and an optimization component. The approximation component can generate an approximate decomposition of a quantum gate. The budget component can set a budget value (C) for a C-factor that is a metric for increase in variance of quasi-probability sampling. The optimization component can determine an optimal decomposition for the quantum gate as a function of C. 1. A system , comprising:a processor that executes the following computer executable components stored in memory, wherein the computer executable components comprise:an approximation component that generates an approximate decomposition of a quantum gate;{'sub': 'budget', 'a budget component that sets, based on the approximate decomposition, a budget value (C) for a C-factor that is a metric for increase in variance of quasi-probability sampling; and'}an optimization component that determines an optimal decomposition for the quantum gate as a function of the C-factor budget.2. The system of claim 1 , wherein the approximation component generates the approximate decomposition utilizing the following equation: [U]≈Σaε claim 1 , wherein:U denotes a unitary corresponding to the quantum gate;M denotes decomposition size;{'sub': 'i', 'adenotes quasi-probability coefficients; and'}{'sub': 'i', 'εdenotes quantum channels implementable on quantum hardware.'}3. The system of claim 2 , wherein the C-factor=C(a claim 2 , . . . claim 2 , a):=Σ|a|.4. The system of claim 3 , wherein the optimization component minimizes ϵ(a claim 3 , . . . claim 3 , a) such that C(a claim 3 , . . . claim 3 , a)≤C claim 3 , and wherein ϵ denotes error.5. The system of claim 1 , further comprising a distribution component that distributes the C-factor across N number of ...

TECHNIQUES FOR SERVICE EXECUTION AND MONITORING FOR RUN-TIME SERVICE COMPOSITION

A server system may receive two or more Quality of Service (QoS) dimensions for the multi-objective optimization model, wherein the two or more QoS dimensions include at least a first QoS dimension and a second QoS dimension. The server system may maximize the multi-objective optimization model along the first QoS dimension, wherein the maximizing includes selecting one or more pipelines for the multi-objective optimization model in the software architecture that meet QoS expectations specified for the first QoS dimension and the second QoS dimension, wherein an ordering of the pipelines is dependent on which QoS dimensions were optimized and de-optimized and to what extent, wherein the multi-objective optimization model is partially de-optimized along the second QoS dimension in order to comply with the QoS expectations for the first QoS dimension, and whereby there is a tradeoff between the first QoS dimension and the second QoS dimension. 1. A method for automating a run-time adaption of a multi-objective optimization model in a software architecture , the method comprising:receiving two or more Quality of Service (QoS) dimensions for the multi-objective optimization model, wherein the two or more QoS dimensions include at least a first QoS dimension and a second QoS dimension; and wherein the maximizing includes selecting one or more pipelines for the multi-objective optimization model in the software architecture that meet QoS expectations specified for the first QoS dimension and the second QoS dimension,', 'wherein an ordering of the pipelines is dependent on which QoS dimensions were optimized and de-optimized and to what extent,', 'wherein the multi-objective optimization model is partially de-optimized along the second QoS dimension in order to comply with the QoS expectations for the first QoS dimension, and', 'whereby there is a tradeoff between the first QoS dimension and the second QoS dimension., 'maximizing the multi-objective optimization model ...

Tissue Staining Quality Determination

The invention relates to the automated determination of the staining quality of an IHC stained biological sample. A plurality of features is extracted from a digital IHC stained tissue image. The features are input into a first classifier configured to identify the extended tissue type of the depicted tissue as a function of the extracted features. An extended tissue type is a tissue type with a defined expression level of the tumor marker. In addition, the extracted features are input into a second classifier configured to identify a contrast level of the depicted tissue as a function of at least some second ones of the extracted features. The contrast level indicates the intensity contrast of pixels of the stained tissue. Then, a staining quality score of the image is computed as a function of the identified extended tissue type and the identified contrast level. 1. An image analysis method for automatically determining the staining quality of an IHC stained biological sample , the method comprising:receiving a digital image of an IHC stained tissue sample of a patient, the pixel intensities of the image correlating with the amount of a tumor-marker-specific stain;extracting a plurality of features from the received digital image;inputting the extracted features into a first classifier, the first classifier being configured to identify the extended tissue type of the tissue depicted in the digital image as a function of at least some first ones of the extracted features, the extended tissue type being a tissue type with a defined expression level of the tumor marker;inputting the extracted features into a second classifier, the second classifier being configured to identify a contrast level of the tissue depicted in the digital image as a function of at least some second ones of the extracted features, the contrast level indicating the intensity contrast of pixels of the stained tissue;computing a staining quality score for the tissue depicted in the digital image ...

Detecting a User's Outlier Days Using Data Sensed by the User's Electronic Devices

A method for detecting a user's outlier days uses data corresponding to features of the user acquired over multiple days by sensors on the user's electronic device. The data acquired for each day and feature is labeled as regular or irregular by applying N labeling approaches. One of the N labeling approaches compares the data for each feature with how values of previously acquired data for corresponding features are distributed. N labels are generated for the data for each feature and day. The machine learning classification model is trained using one of the N labels for each of the N labeling approaches. An optimal labeling approach is selected from among the N labeling approaches for each feature using the machine learning classification model. For each feature, the method determines whether each of the days is an outlier day for the user using the labels obtained with the optimal labeling approach. 133-. (canceled)34. A method for detecting an outlier time period of a user , comprising:receiving data corresponding to each of a plurality of features related to the user, wherein the data is acquired over a plurality of time periods by sensors on an electronic device of the user;labeling the data acquired on each of the plurality of time periods for each of the plurality of features as being regular or irregular by applying N labeling approaches, wherein N is a positive integer greater than one, and wherein at least one of the N labeling approaches involves comparing the data corresponding to each of the plurality of features with how values of previously acquired data for corresponding features are distributed;generating N labels for the data corresponding to each of the plurality of features for each of the plurality of time periods;for each of the N labeling approaches, using one of the N labels to train a machine learning classification model;selecting from among the N labeling approaches an optimal labeling approach for each of the plurality of features using ...

MULTI-MODAL REPRESENTATION BASED EVENT LOCALIZATION

A method performed by an artificial neural network (ANN) includes determining, at a first stage of a multi-stage cross-attention model of the ANN, a first cross-correlation between a first representation of each modality of a number of modalities associated with a sequence of inputs. The method still further includes determining, at each second stage of one or more second stages of the multi-stage cross-attention model, a second cross-correlation between first attended representations of each modality. The method also includes generating a concatenated feature representation associated with a final second stage of the one or more second stages based on the second cross-correlation associated with the final second stage, the first attended representation of each modality, and the first representation of each modality. The method further includes determining a probability distribution between a set of background actions and a set of foreground actions from the concatenated feature representation. The method still further includes localizing an action in the sequence of inputs based on the probability distribution. 1. A method performed by an artificial neural network (ANN) , comprising:determining, at a first stage of a multi-stage cross-attention model of the ANN, a first cross-correlation between a first representation of each modality of a plurality of modalities associated with a sequence of inputs;{'claim-text': ['a first attended representation of a first modality of the plurality of modalities based on the first cross-correlation and the first representation of the first modality, and', 'the first attended representation of a second modality of the plurality of modalities based on the first cross-correlation and the first representation of the second modality;'], '#text': 'determining, at each second stage of one or more second stages of the multi-stage cross-attention model, a second cross-correlation between first attended representations of each modality,'} ...

REPLACING STAIR-STEPPED VALUES IN TIME-SERIES SENSOR SIGNALS WITH INFERENTIAL VALUES TO FACILITATE PROGNOSTIC-SURVEILLANCE OPERATIONS

During operation, the system obtains the time-series sensor signals, which were gathered from sensors in a monitored system. Next, the system classifies the time-series sensor signals into stair-stepped signals and un-stair-stepped signals. The system then replaces stair-stepped values in the stair-stepped signals with interpolated values determined from un-stair-stepped values in the stair-stepped signals. Next, the system divides the time-series sensor data into a training set and an estimation set. The system then trains an inferential model on the training set, and uses the trained inferential model to replace interpolated values in the estimation set with inferential estimates. Next, the system switches roles of the training and estimation sets to produce a new training set and a new estimation set. The system then trains the inferential model on the new training set, and uses the trained inferential model to replace interpolated values in the new estimation set with inferential estimates. 1. A method for preprocessing time-series sensor signals to facilitate prognostic-surveillance operations , comprising:obtaining the time-series sensor signals, which were gathered from sensors in a monitored system during operation of the monitored system;classifying the time-series sensor signals into stair-stepped signals and un-stair-stepped signals;performing an interpolation operation to replace stair-stepped values in the stair-stepped signals with interpolated values determined from un-stair-stepped values in the stair-stepped signals; andusing an inferential model to replace the interpolated values with inferential estimates determined based on correlations among the time-series sensor signals.2. The method of claim 1 , wherein using an inferential model to replace the interpolated values with the inferential estimates comprises:dividing the time-series sensor data into a training set and an estimation set;training an inferential model on the training set;using the ...

METHOD AND SYSTEM OF MATCHING DATA IN USING MULTIPLE DATA POINTS

A method for matching supplied organizational data with trade directory information includes: receiving a data file including a plurality of organizational entries, each entry including an entity name and a geographic location; normalizing the entity name in each of the organizational entries; identifying a plurality of matching entries for each organizational entries, each matching entry including a matching name and location, and where each matching entry is identified based on a first correspondence between the matching name and the entity name and a second correspondence between the matching location and the geographic location; determining a confidence level for each of the organizational entries based on the correspondence between the first correspondence and the second correspondence for at least one of the identified matching entries; and transmitting the plurality of matching entries and determined confidence level for each of the plurality of organizational entries. 1. A method for matching supplied organizational data with trade directory information , comprising:receiving, by a receiver of a processing server, a data file including a plurality of organizational entries from a computing system, each organizational entry including at least an entity name and a geographic location;normalizing, by a processor of the processing server, the entity name included in each of the plurality of organizational entries;identifying, by the processor of the processing server, a plurality of matching entries for each of the plurality of organizational entries, wherein each matching entry includes at least a matching name and a matching location, and where each matching entry is identified based on a first correspondence between the matching name and the respective normalized entity name and a second correspondence between the matching location and the respective geographic location;determining, by the processor of the processing server, a confidence level for each of the ...

Loss Function Optimization Using Taylor Series Expansion

A process for optimizing loss functions includes progressively building better sets of parameters for loss functions represented as multivariate Taylor expansions in accordance with an iterative process. The optimization process is built upon CMA-ES. At each generation (i.e., each CMA-ES iteration), a new set of candidate parameter vectors is sampled. These candidate parameter vectors are sampled from a multivariate Gaussian distribution representation that is modeled by the CMA-ES covariance matrix and the current mean vector. The candidates are then each evaluated by training a model (neural network) using the candidates and determining a fitness value for each candidate against a validation data set. 1. A process for optimizing a loss function for a model solving a particular problem comprising:providing an initial mean solution vector to a multi-dimensional continuous value optimization process running on one or more processors;(ii) generating a set of candidate loss function parameters using the initial mean solution vector for use in building a first set of candidate loss functions in accordance with a predetermined loss function representation; (a) building each of the first candidate loss functions using the initial set of candidate loss function parameters;', '(b) at least partially training the model on a training data set related to the particular problem using each of the first candidate loss functions;', '(c) evaluating the model trained with each of the first candidate loss functions on a validation data set related to the particular problem;', '(d) obtaining individual fitness values for each of the first candidate loss functions from the evaluation in (c);, '(iii) evaluating each of the candidate loss functions in the first set of candidate loss function with the model including(iv) ranking each of the first candidate loss functions in accordance with individual fitness values, wherein each of the first candidate loss functions includes a different ...

QUANTIZED INPUTS FOR MACHINE LEARNING MODELS

Methods and apparatus for pre-processing first data for use with a trained machine learning model. In some embodiments, the method may comprise accessing the first data, wherein the first data has a first precision; generating, based on at least a first portion of the first data, second data having a second precision lower than the first precision; and providing the second data as input to the trained machine learning model to generate model output. 1. A method of pre-processing first data for use with a trained machine learning model , the method comprising:accessing the first data, wherein the first data has a first precision;generating, based on at least a first portion of the first data, second data having a second precision lower than the first precision; andproviding the second data as input to the trained machine learning model to generate model output.2. The method of claim 1 , wherein the first precision comprises a first number of bits claim 1 , and the second precision comprises a second number of bits claim 1 , wherein the second number of bits is lower than the first number of bits.3. The method of claim 1 , wherein the trained machine learning model was trained using training data having a same precision as the first precision.4. The method of claim 1 , wherein the at least the first portion of the first data comprises all of the first data.5. The method of claim 1 , further comprising scaling and/or normalizing the second data.6. The method of claim 1 , wherein generating the second data is performed as part of an analog to digital conversion of the at least the first portion of the first data.7. The method of claim 2 , wherein the second number of bits is less than 65% of the first number of bits claim 2 , and wherein an accuracy of the trained machine learning model using the second data is at least 95% of an accuracy of the trained machine learning model using data having a same precision as the first precision.8. The method of claim 1 , wherein ...

IDENTIFYING A PROCESS AND GENERATING A PROCESS DIAGRAM

A device may receive activity data identifying activities of a process performed by users via user devices. The device may receive baseline data identifying baselines associated with the process and variant data identifying variants from the baselines. The device may apply a sequence alignment model, to the activity data and the baseline data, to determine first similar sequences associated with the activities and the baselines and may apply the sequence alignment model, to the activity data and the variant data, to determine second similar sequences associated with the activities and the variants. The device may determine, based on the first similar sequences, first label data identifying first labels for the activities and may determine, based on the second similar sequences, second label data identifying second labels for the activities. The device may generate a process diagram based on the activity data, the first label data, and the second label data.

METHODS AND SYSTEMS FOR TRAINING CONVOLUTIONAL NEURAL NETWORKS

A computer implemented method and system for training a convolutional neural network is provided. The method includes receiving a captured image. Based on the captured image, a statistical noise model is generated. A convolutional neural network is trained based on the captured image and the statistical model.

TECHNIQUES FOR AUTOMATED DATA CLEANSING FOR MACHINE LEARNING ALGORITHMS

Machine learning models typically are based on processing large-volume datasets, and datasets are preprocessed so that the machine learning can provide sound results. In building a model, certain example embodiments generate meta-features for each of a number of independent variables in an accessed portion of the dataset. The meta-features are provided as input to pre-trained classification models. Those models output, for the independent variables, indications of one or more appropriate missing value imputation operations, and one or more appropriate other preprocessing data cleansing related operations. The data in the dataset is transformed by selectively applying the missing value imputation operation(s) and the other preprocessing operation(s), in accordance with the independent variables associated with the data, thereby performing the preprocessing in an automated and programmatic way that helps improve the quality of the built model. Ultimately, queries received over a computer-mediated interface can be processed using the built machine learning model. 1. A machine learning system , comprising:a non-transitory computer readable storage medium storing thereon a dataset having data from which a machine learning model is buildable;an electronic computer-mediated interface configured to receive a query processable in connection with a machine learning model; accessing at least a portion of the dataset;', generating meta-features for the respective independent variable;', 'providing, as input to at least first and second pre-trained classification models that are different from one another, the generated meta-features for the respective independent variable;', 'receiving, as output from the first pre-trained classification model, an indication of one or more missing value imputation operations appropriate for the respective independent variable; and', 'receiving, as output from the second pre-trained classification model, an indication of one or more other ...

MULTI-TASK MULTI-MODAL MACHINE LEARNING SYSTEM

Methods, systems, and apparatus, including computer programs encoded on computer storage media for training a machine learning model to perform multiple machine learning tasks from multiple machine learning domains. One system includes a machine learning model that includes multiple input modality neural networks corresponding to respective different modalities and being configured to map received data inputs of the corresponding modality to mapped data inputs from a unified representation space; an encoder neural network configured to process mapped data inputs from the unified representation space to generate respective encoder data outputs; a decoder neural network configured to process encoder data outputs to generate respective decoder data outputs from the unified representation space; and multiple output modality neural networks corresponding to respective different modalities and being configured to map decoder data outputs to data outputs of the corresponding modality. 1. A system comprising one or more computers and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to implement: a plurality of input modality neural networks, wherein each input modality neural network corresponds to a different modality of multiple modalities and is configured to map received data inputs of the corresponding modality to mapped data inputs from a unified representation space;', 'an encoder neural network that is configured to process mapped data inputs from the unified representation space to generate respective encoder data outputs;', 'a decoder neural network that is configured to process encoder data outputs to generate respective decoder data outputs from the unified representation space; and', 'a plurality of multiple output modality neural networks, wherein each output modality neural network corresponds to a different modality and is configured to map decoder data outputs from the unified ...

COMPUTER ARCHITECTURE FOR MAPPING ANALOG DATA VALUES TO A STRING CORRELITHM OBJECT IN A CORRELITHM OBJECT PROCESSING SYSTEM

A string correlithm object generator is configured to output a string correlithm object comprising a plurality of sub-string correlithm objects. A node is configured to receive a plurality of data values. A memory is configured to store a node table that associates sub-string correlithm objects with the data values such that a first sub-string correlithm object is associated with a first data value and a second sub-string correlithm object is associated with a second data value. A processor is configured to receive a third data value that is between the first data value and the second data value, determine a third sub-string correlithm object that is interpolated between the first sub-string correlithm object and the second sub-string correlithm object, and associate the third sub-string correlithm object with the third data value. 1. A device configured to emulate a string correlithm object in a correlithm object processing system , comprising:a string correlithm object generator configured to output a string correlithm object comprising a plurality of sub-string correlithm objects, wherein each sub-string correlithm object is adjacent in n-dimensional space to a preceding sub-string correlithm object and a subsequent sub-string correlithm object to form a string;a node configured to receive a plurality of data values; the node table associates sub-string correlithm objects with the data values such that a first sub-string correlithm object is associated with a first data value and a second sub-string correlithm object is associated with a second data value;', 'the first sub-string correlithm object is represented by a first n-bit digital word; and', 'the second sub-string correlithm object is represented by a second n-bit digital word that differs from the first n-bit digital word by a predetermined number of bits;, 'a memory configured to store a node table, wherein receive a third data value that is between the first data value and the second data value;', ' ...

MISSING DATA COMPENSATION METHOD, MISSING DATA COMPENSATION SYSTEM, AND NON-TRANSITORY COMPUTER-READABLE MEDIUM

A missing data compensation method, missing data compensation system and non-transitory computer-readable medium are provided in this disclosure. The method includes the following operations: inputting a sensing signal by a sensor; searching for a historical data sections similar to a first data section from the plurality of historical data sections to generate a plurality of candidate data sections; calculating a plurality of data relation diagrams according to the first data section and the candidate data sections, respectively; utilizing a feature recognition model to calculate a plurality of similarity values according to the data relation diagrams; selecting a candidate data section corresponding to the maximum similarity value as a sample data section; and utilizing the data in the sample data section to compensate the data in the first data section to generate compensated data section. 1. A missing data compensation method , comprising:inputting a sensing signal by a sensor, wherein the sensing signal comprises a plurality of data sections, and a historical database comprises a plurality of historical data sections;searching for a historical data sections similar to a first data section from the plurality of historical data sections to generate a plurality of candidate data sections, wherein the first data section is one of the plurality of data sections;calculating a plurality of data relation diagrams according to the first data section and the plurality of candidate data sections, respectively;utilizing a feature recognition model to calculate a plurality of similarity values according to the plurality of data relation diagrams, respectively and selecting a candidate data section corresponding to maximum similarity value as a sample data section; andutilizing data in the sample data section to compensate data in the first data section to generate compensated data section.2. The missing data compensation method of claim 1 , further comprising:inputting a ...

MODEL STRUCTURE SELECTION APPARATUS, METHOD, DISAGGREGATION SYSTEM AND PROGRAM

Provided an apparatus that receives time series data from a data storage unit storing time series of sample data or feature values calculated from the sample data, computes a measure indicating change and repetition characteristics of the time series data, based on sample value distribution thereof, selects a state model structure to be used for model learning and estimation, from state models including a fully connected state model and a one way direction state model, based on the measure and stores the selected state model in a model storage unit. 1. A state model structure selection apparatus comprising:a processor; anda memory storing program instructions executable by the processor is configured to execute the program instructions stored in the memory toreceive time series data from a data storage unit that stores time series of sample data or feature values calculated from the sample data;compute a measure indicating change and repetition characteristics of the time series data, based on sample value distribution thereof; andselect a state model structure to be used for model learning and estimation, from state models including a fully connected state model and a one way direction state model, based on the measure and stores the selected state model in a storage unit that stores the state model selected.2. The state model structure selection apparatus according to claim 1 , wherein the processor is configured to compute:a first probability of number of cycles for each magnitude value in the time series data, by dividing number of cycles of the magnitude value by total number of cycles in the time series data; anda second probability of number of occurrences for each magnitude value in the time series data, by dividing number of occurrences of the magnitude value by a length of the time series data;compute a correlation coefficient between the first probability and the second probability, andselect either the fully connected state model or the one way direction ...

Processor, and method for generating binarized weights for a neural network

A processor for generating binarized weights for a neural network. The processor comprises a binarization scheme generation module configured to generate, for a group of weights taken from a set of input weights for one or more layers of a neural network, one or more potential binary weight strings representing said group of weights; a binarization scheme selection module configured to select a binary weight string to represent said group of weights, from among the one or more potential binary weight strings, based at least in part on a number of data bits required to represent the one or more potential binary weight strings according to a predetermined encoding method; and a weight generation module configured to output data representing the selected binary weight string for representing the group of weights.

STORAGE MEDIUM, SHAPE DATA OUTPUT METHOD, AND INFORMATION PROCESSING DEVICE

A non-transitory computer-readable storage medium storing a shape data output program that causes at least one computer to execute process, the process includes, normalizing each shape data of a plurality of pieces of shape data for each component in each coordinate axis direction to create unit shape data; classifying the plurality of pieces of shape data based on the created unit shape data of each of the pieces of shape data; specifying, based on dimensions of sites of each shape data in classified group, a dimensional relationship between different sites of the shape data in the group; and outputting information indicating the specified dimensional relationship in association with the unit shape data of the shape data in the group. 1. A non-transitory computer-readable storage medium storing a shape data output program that causes at least one computer to execute process , the process comprising:normalizing each shape data of a plurality of pieces of shape data for each component in each coordinate axis direction to create unit shape data;classifying the plurality of pieces of shape data based on the created unit shape data of each of the pieces of shape data;specifying, based on dimensions of sites of each shape data in classified group, a dimensional relationship between different sites of the shape data in the group; andoutputting information indicating the specified dimensional relationship in association with the unit shape data of the shape data in the group.2. The non-transitory computer-readable storage medium according to claim 1 , whereinthe creating includesextracting a minimum value in each coordinate axis direction from coordinates of each characteristic point of each of the pieces of shape data,subtracting the extracted minimum value in each coordinate axis direction from each value of the coordinates of each characteristic point,extracting a maximum value in each coordinate axis direction from the coordinates of each characteristic point after ...

ARTIFICIAL INTELLIGENCE-BASED TEXT-TO-SPEECH SYSTEM AND METHOD

A technique improves training and speech quality of a text-to-speech (TTS) system having an artificial intelligence, such as a neural network. The TTS system is organized as a front-end subsystem and a back-end subsystem. The front-end subsystem is configured to provide analysis and conversion of text into input vectors, each having at least a base frequency, f, a phenome duration, and a phoneme sequence that is processed by a signal generation unit of the back-end subsystem. The signal generation unit includes the neural network interacting with a pre-existing knowledgebase of phenomes to generate audible speech from the input vectors. The technique applies an error signal from the neural network to correct imperfections of the pre-existing knowledgebase of phenomes to generate audible speech signals. A back-end training system is configured to train the signal generation unit by applying psychoacoustic principles to improve quality of the generated audible speech signals. 1. A text-to-speech (TTS) system including one or more processors and one or more memories configured to perform operations for converting text into a corrected speech signal comprising:interacting with data of previously generated speech in a pre-existing knowledgebase of phonemes, wherein the previously generated speech has speech signal distortions;generating the corrected speech signal of the previously generated speech to correct for the speech signal distortions of the previously generated speech based upon, at least in part, interacting with the data of the previously generated speech in the pre-existing knowledgebase of phonemes; andapplying the corrected speech signal to the previously generated speech for correcting the speech signal distortions of the previously generated speech in the pre-existing knowledgebase of phonemes.2. The TTS system of wherein the operations further comprise converting a frequency domain signal combined from a neural network and the pre-existing knowledgebase ...

ARTIFICIAL NEURAL NETWORK COMPUTING SYSTEMS

The present disclosure relates to an artificial neural network (ANN) computing system comprising: a buffer configured to store data indicative of input data received from an input device; an inference engine operative to process data from the buffer to generate an interest metric for the input data; and a controller. The controller is operative to control a mode of operation of the inference engine according to the interest metric for the input data. 1. An artificial neural network (ANN) computing system comprising:a buffer configured to store data indicative of input data received from an input device;an inference engine operative to process data from the buffer to generate an interest metric for the input data; anda controller,wherein the controller is operative to control a mode of operation of the inference engine according to the interest metric for the input data.2. An ANN system according to claim 1 , wherein the controller is operative to issue a first control signal to adjust the mode of operation of the inference engine if the interest metric exceeds a threshold.3. An ANN system according to wherein the inference engine is operable in a first mode of operation and a second mode of operation.4. An ANN system according to wherein in the first mode of operation the inference engine is operative to generate the interest metric based on data from the buffer associated with a particular point or period in time.5. An ANN system according to wherein the controller is further operative to issue a second control signal to cause the inference engine to process data from the buffer associated with a predetermined period of time claim 2 , prior to a point or period in time associated with the input data for which the interest metric was generated claim 2 , if the interest metric exceeds the threshold.6. An ANN system according to wherein the inference engine implements a multi-layer artificial neural network and wherein:in the first mode of operation the inference ...

METHOD FOR PREDICTING A DISEASE OUTBREAK

The present invention relates to a method for predicting a disease outbreak. This method includes the steps of collecting data by the remote data interface, validating the obtained data by the remote data interface, analysing data and computing new parameters by a data analysis interface, identifying the association between the case, the outbreak and various indicators, and establishing one-to-one relationships between each case and the index case of the outbreak, and one-to-many relationships between each case and other cases of the outbreak by the remote data interface, predicting case parameters by a data prediction interface, wherein the predicted case parameters include time of the outbreak and number of people affected by the outbreak, predicting severity of the outbreak by the data prediction interface, and predicting geographical location of an epicentre of the outbreak by the data prediction interface. 1. A method for predicting parameters of a disease outbreak is characterised by the steps of:a) collecting data regarding a disease by a remote data interface, the data includes clinical data from health centres and environmental data from various sources;b) validating the obtained data by the remote data interface;c) analysing data and computing new parameters by a data analysis interface;d) identifying associations and relationships among the disease cases, the outbreak and various indicators by the remote data interface;e) predicting case parameters by a data prediction interface, wherein the predicted case parameters include time of the outbreak and number of people affected by the outbreak;f) predicting severity of the outbreak by the data prediction interface; andg) predicting geographical location of an epicentre of the outbreak by the data prediction interface.2. The method as claimed in claim 1 , wherein collecting data regarding a disease by the remote data interface includes the steps of:a) determining whether clinical data is available from health ...

HUMAN MONITORING SYSTEM INCORPORATING CALIBRATION METHODOLOGY

Related methods are provided for establishing a baseline value to represent an eyelid opening dimension for a person engaged in an activity, where the activity may be driving a vehicle, operating industrial equipment, or performing a monitoring or control function; and for operating a system for monitoring eyelid opening values with real time video data. 1. A method for establishing a baseline value to represent an eyelid opening dimension for a person engaged in an activity , where the activity may be driving a vehicle , operating industrial equipment , or performing a monitoring or control function , comprising:while the person is engaged in performing the activity, serially acquiring a temporal sequence of image frames, with each in a plurality of the image frames containing eyelid opening data for the person;using eyelid opening data derived from the plurality of image frames, serially processing each in the plurality of image frames when acquired to sequentially calculate members of a data set comprising a measured eyelid opening value for each in the plurality of image frames;while sequentially calculating members of the data set, computing an interim measure, as a first baseline eyelid opening value, based on a first portion of the total number of eyelid opening values being calculated, which measure is representative of eyelid opening values in a first portion of the plurality of image frames; andafter computing the interim measure, computing a second measure based on a different portion of the eyelid opening values which different portion may include some or all of the first portion of the eyelid opening values and does include one or more eyelid opening values derived from image frames acquired after acquisition of the first portion of the plurality of image frames.2. The method of wherein claim 1 , when computing the second measure claim 1 , said different portion of the eyelid opening values does include at least some of the first portion of the eyelid ...

HUMAN MONITORING SYSTEM INCORPORATING CALIBRATION METHODOLOGY

Method for monitoring eyelid opening values. In one embodiment video image data is acquired with a camera which data are representative of a person engaged in an activity, where the activity may be driving a vehicle, operating industrial equipment, or performing a monitoring or control function. When the person's head undergoes a change in yaw angle, such that eyelids of both eyes of the person are captured with the camera, but one eye is closer to the camera than the other eye, a weighting factor is applied, which factor varies as a function of the yaw angle such that a value representative of eyelid opening data based on both eyes is calculated. 1. A method for monitoring eyelid opening values by acquiring video image data with a camera which data are representative of a person engaged in an activity , where the activity may be driving a vehicle , operating industrial equipment , or performing a monitoring or control function , comprising: {'br': None, 'i': w', 'w, '(LEOD)+(1−)(REOD), where, "when the person's head undergoes a change in yaw angle, such that eyelids of both eyes of the person are captured with the camera, but one eye is closer to the camera than the other eye, applying a weighting factor which varies as a function of the yaw angle such that a value representative of eyelid opening data based on both eyes is calculated as:"}LEOD is the measured left eyelid opening distance and REOD is the measured right eyelid opening distance; andwhere the weight, w, varies from zero to one and changes proportionally to the change in head yaw angle.2. The method of where the weight claim 1 , w claim 1 , varies from zero to one and changes proportionally to the change in head yaw angle between a preselected range of angles [−φ claim 1 , +φ].3. The method of where the preselected range extends between φ=−15° to φ=15°.4. The method of where the preselected range extends between φ=−15° to φ=15°.5. The method of where w=(0.5)(θ+30)/30 where θ is the yaw angle.6. The ...

NEURAL NETWORK RECOGNTION AND TRAINING METHOD AND APPARATUS

Disclosed is a recognition and training method and apparatus. The apparatus may include a processor configured to input data to a neural network, determine corresponding to a multiclass output a mapping function of a first class and a mapping function of a second class, acquire a result of a loss function including a first probability component that changes correspondingly to a function value of the mapping function of the first class and a second probability component that changes contrastingly to a function value of the mapping function of the second class, determine a gradient of loss corresponding to the input data based on the result of the loss function, update a parameter of the neural network based on the determined gradient of loss for generating a trained neural network based on the updated parameter. The apparatus may input other data to the trained neural network, and indicate a recognition result. 1. A processor implemented training method comprising:inputting input data to a neural network;determining respective mapping functions corresponding to a multiclass output of the neural network in association with the input data, including determining a mapping function of a first class and a mapping function of a second class;acquiring a result of a loss function including a first probability component that changes correspondingly to a function value of the mapping function of the first class and a second probability component that changes contrastingly to a function value of the mapping function of the second class;determining a gradient of loss corresponding to the input data based on the result of the loss function; andupdating a parameter of the neural network based on the determined gradient of loss for generating a trained neural network based on the updated parameter.2. The method of claim 1 , wherein the first probability component increases with respect to increases in the function value of the mapping function of the first class and the second ...

MEASURING CONFIDENCE IN DEEP NEURAL NETWORKS

A system comprises a computer including a processor and a memory, and the memory including instructions such that the processor is programmed to calculate a standard deviation of a plurality of predictions, wherein each prediction of the plurality of predictions is generated by a different deep neural network using sensor data; and determine at least one of a measurement corresponding to an object based on the standard deviation. 1. A system comprising a computer including a processor and a memory , the memory including instructions such that the processor is programmed to:calculate a standard deviation of a plurality of predictions, wherein each prediction of the plurality of predictions is generated by a different deep neural network using sensor data; anddetermine at least one of a measurement corresponding to an object based on the standard deviation.2. The system of claim 1 , wherein the processor is further programmed to:compare the standard deviation of a distribution with a predetermined variation threshold; andtransmit, to a server, the sensor data when the standard deviation is greater than the predetermined variation threshold.3. The system of claim 2 , wherein the process is further programmed to:disable an autonomous vehicle mode of a vehicle when the standard deviation is greater than a predetermined distribution variation threshold.4. The system of claim 1 , wherein the processor is further programmed to:receive the sensor data from a vehicle sensor of a vehicle; andprovide the sensor data to each deep neural network.5. The system of claim 1 , wherein each deep neural network comprises a convolutional neural network.6. The system of claim 5 , wherein the processor is further programmed to:provide an image captured by an image sensor of a vehicle to each convolutional neural network; andcalculate the plurality of predictions based on the image.7. The system of claim 1 , wherein the object comprises at least a portion of a trailer connected to a vehicle ...

METHOD FOR TRAINING NEURAL NETWORK

Disclosed is a computer program stored in a computer readable storage medium, in which when the computer program is executed in one or more processors, the computer program performs operations for training a neural network, the operations including: displaying a first screen including at least one first object receiving a selection input for a project; and displaying a second screen for displaying information related to the project corresponding to the selected project, in which the second screen includes at least one of a first output portion for displaying time series data obtained from a sensor or a second output portion for displaying a selection portion including at least one second object for receiving a selection input related to a model retraining or information corresponding to the second object. 1. A computer program stored in a computer readable medium , wherein when the computer program is executed by one or more processors of a computing device , the computer program performs operations to provide methods for training neural networks , and the operations comprise:displaying, by a processor, a first screen including at least one first object receiving a selection input for a project; anddisplaying, by the processor, a second screen including information related to the project corresponding to the selected project,wherein the second screen includes at least one of a first output portion for displaying time series data obtained from a sensor or a second output portion for displaying a selection portion including at least one second object receiving a selection input for a model retraining or information corresponding to the second object.2. The computer program according to claim 1 , wherein the project is a project related to an artificial intelligence for achieving a specific goal based on the artificial intelligence claim 1 , andthe specific goal includes the goal of improving the performance of the model applied the artificial intelligence.3. The ...

DYNAMIC OUTLIER BIAS REDUCTION SYSTEM AND METHOD

In at least one embodiment, the present description is directed to a computer system, having at least components of a server, including a processor and a non-transient storage subsystem, storing a computer program including instructions that, when executed by the processor, cause the processor to at least: electronically receive a model for one or more operating conditions, one or more threshold criteria, and facility operating data for each respective facility of a plurality of facilities; validate the one or more threshold criteria to be one or more acceptable bias criteria; iteratively perform one or more iterations of outlier bias reduction in the facility operating data based on the model; determine, based on non-biased facility operating data, a non-biased performance standard for the one or more operating conditions; and track, based on the non-biased performance standard and the facility operating data, operating performance of each respective facility of the plurality of facilities. 1. A method comprising the steps of: i) a model for one or more operating conditions,', 'ii) one or more threshold criteria, and', 'iii) facility operating data of the one or more operating conditions for each respective facility of a plurality of facilities;, 'electronically receiving, by a processor, at least the followingwherein the model comprises one or more coefficients;validating, by the processor, the one or more threshold criteria to be one or more acceptable bias criteria;iteratively performing, by the processor, one or more iterations of outlier bias reduction in the facility operating data of the plurality of facilities based at least in part on the model; (i) determining a set of model predicted values;', '(ii) comparing the set of model predicted values to the facility operating data to produce a set of error values;', (iv) constructing, based at least in part on the non-biased facility operating a data, an updated model for the one or more operating conditions, ...

NON-LOCAL MEANS DENOISING

Systems and methods are disclosed for non-local means denoising of images. For example, methods may include receiving an image from an image sensor; determining a set of non-local means weights for the image; applying the set of non-local means weights to the image to obtain a first denoised image; applying the set of non-local means weights to the first denoised image to obtain a second denoised image; and storing, displaying, or transmitting an output image based on the second denoised image. 1. A system comprising:an image sensor configured to capture an image; and receive the image from the image sensor;', 'determine a set of weights for the image, wherein a weight in the set of weights corresponds to a subject pixel and a candidate pixel and is determined based on values of one or more pixels of the image centered at the subject pixel and one or more pixels of the image centered at the candidate pixel;', 'apply the set of weights to the image to obtain a first denoised image, wherein the subject pixel of the first denoised image is determined based on the weight multiplied by the candidate pixel of the image;', 'apply the set of weights to the first denoised image to obtain a second denoised image, wherein the subject pixel of the second denoised image is determined based on the weight multiplied by the candidate pixel of the first denoised image; and', 'store, display, or transmit an output image based on the second denoised image., 'a processing apparatus configured to2. The system of claim 1 , in which claim 1 , for each pixel of the image there is a subset of the set of weights that corresponds to that pixel as subject claim 1 , and the subset includes weights respectively associated with a plurality of other pixels that are candidates for that pixel.3. The system of claim 2 , in which the candidates for the subject pixel are pixels of the image in an area centered at the subject pixel.4. The system of claim 1 , in which the processing apparatus is ...

MEDICAL IMAGE PROCESSING APPARATUS AND MEDICAL IMAGE PROCESSING SYSTEM

A medical image processing apparatus according to an embodiment comprises a memory and processing circuitry. The memory is configured to store a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site. The processing circuitry is configured to process first data into second data using, among the neural networks, the neural network corresponding to the imaging target site for the first data, wherein the first data is input to the input layer and the second data is output from the output layer. 1. A medical image processing apparatus comprising:a memory configured to store a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site; andprocessing circuitry configured to process first data into second data using, among the neural networks, the neural network corresponding to the imaging target site for the first data, wherein the first data is input to the input layer and the second data is output from the output layer.2. The medical image processing apparatus according to claim 1 , whereinthe memory is configured to store a plurality of neural networks corresponding to a plurality of imaging conditions, respectively, the neural networks corresponding to the respective imaging conditions each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets ...

Gaze point estimation method, device, and electronic device

The present application provides a gaze point estimation method, device, and an electronic device. The method includes: acquiring user image data; acquiring a facial feature vector according to a preset first convolutional neural network and the facial image; acquiring a position feature vector according to a preset first fully connected network and the position data; acquiring a binocular fusion feature vector according to a preset eye feature fusion network, the left-eye image and the right-eye image; and acquiring position information about a gaze point of a user according to a preset second fully connected network, the facial feature vector, the position feature vector, and the binocular fusion feature vector. In this technical solution, relation between eye images and face images is utilized to achieve accurate gaze point estimation.

BEHAVIOR CLASSIFICATION AND PREDICTION THROUGH TEMPORAL FINANCIAL FEATURE PROCESSING WITH RECURRENT NEURAL NETWORK

A system, computer program product, and method are presented for classifying behaviors and predictions through processing temporal financial features with a recurrent neural network (RNN). The method includes receiving, by a RNN model, first financial transaction events. The method also includes classifying non-fraudulent behavioral patterns and potentially fraudulent behavioral patterns resident within the first financial transaction events and training the RNN model therewith. The method further includes receiving, by the RNN model, second financial transaction events over a predetermined period of time. The method also includes normalizing the second financial transaction events, including partitioning the predetermined period of time into a plurality of first equal temporal segments. Some of the plurality of first equal temporal segments are representative of the second financial transaction events residing therein. The method further includes predicting a labeling of the second financial transaction events with a behavior pattern of one of non-fraudulent and potentially fraudulent. 1. A computer system comprising:one or more processing devices and at least one memory device operably coupled to the one or more processing devices; receive, by the RNN model, for one or more first target focal objects, one or more first sequential series of financial transaction events;', 'determine non-fraudulent and potentially fraudulent financial transactions resident within the one or more first sequential series of financial transaction events;', 'classify at least a first portion of the one or more first sequential series of financial transaction events as a non-fraudulent behavioral pattern;', 'classify at least a second portion of the one or more first sequential series of financial transaction events as a potentially fraudulent behavioral pattern;', 'train the RNN model with the non-fraudulent behavioral pattern and the potentially fraudulent behavioral pattern;', ' ...

Object classification based on decoupling a background from a foreground of an image

Techniques facilitating object classification based on decoupling a background from a foreground of an image are provided. A system can comprise a memory that stores computer executable components and a processor that executes the computer executable components stored in the memory. The computer executable components can comprise a model that is trained on images that comprise respective backgrounds and respective foregrounds that are interleaved. The model can be trained to detect the respective backgrounds with a defined confidence level. The computer executable components can also comprise an extraction component that employs the model to identify a background of a received image based on the defined confidence level and to decouple a foreground object of the received image based on identification of the background of the received image.

METHOD FOR DETECTING DOCUMENT FRAUD

A method for detecting a document fraud is disclosed. A first image of a first document and a second image of a second document are obtained. A procedure of detection of zones sensitive to document frauds are applied in the regions of the first image and of the second image registered on the first image. Each sensitive zone detected is then divided into a plurality of subparts. A measurement of dissimilarity is calculated between corresponding subparts from the first image and the registered second image. It is then determined whether the first document is identical to the second document from measurements of dissimilarity. If the first document is different from the second document, a level of difference is determined between the first and second documents according to a value representing a proportion of different subparts; and a fraud is detected when the level of difference is below a third predetermined threshold. 1. A method for detecting document fraud , wherein the method comprises:obtaining a first image of a first document and a second image of a second document;applying an image registration procedure to the second image so as to register it to the first image, the registration procedure being based on a matching of points of interest identified in the first and second images;applying a procedure of detection of zones sensitive to document frauds in the first image and in the registered second image;dividing each sensitive zone detected into a plurality of subparts and, for each subpart, calculating a signature representing a content of said subpart;for each subpart of the first image, seeking a subpart corresponding spatially in the registered second image and, for each subpart of the first image having a corresponding subpart in the second image, calculating a measurement of local dissimilarity between the corresponding subparts from the signatures;determining that the first and second documents are identical when a measurement of global dissimilarity ...

ANNOTATION METHOD AND DEVICE, AND STORAGE MEDIUM

An annotation method and device and a storage medium are provided. The annotation method includes operations as follows. A first probability value that a first sample image is annotated with an Nth tag when the first sample image is annotated with an Mth tag is determined based on first tag information of a first image set. M and N are unequal and are positive integers. The first probability value is added to second tag information of a second sample image annotated with the Mth tag in a second image set. 1. An annotation method , comprising:determining, based on first tag information of a first image set, a first probability value that a first sample image is annotated with an Nth tag when the first sample image is annotated with an Mth tag, wherein M and N are unequal and are positive integers; andadding the first probability value to second tag information of a second sample image, annotated with the Mth tag, in a second image set.2. The annotation method of claim 1 , further comprising:acquiring a third image set in a target application scenario, wherein the third image set comprises third sample images; andannotating different image features of the third sample images by using different types of tags to obtain the first image set before determining, based on the first tag information of the first image set, the first probability value.3. The annotation method of claim 1 , further comprising:classifying the second sample images in the second image set according to a type of tag to obtain subsets of the second image set, wherein each of the sample images in a same subset is annotated with the Mth tag.4. The annotation method of claim 3 , wherein adding the first probability value to the second tag information of the second sample image annotated with the Mth tag in the second image set comprises:for each of the subsets of the second image set, adding in bulk the first probability value to the second tag information of the second sample images annotated with the ...

TRANSMISSION APPARATUS RECOGNITION APPARATUS, LEARNING APPARATUS, TRANSMISSION APPARATUS RECOGNITION METHOD, AND, LEARNING METHOD

A transmission apparatus recognition apparatus includes a storage unit that stores K sets of template feature groups for estimating K (an integer of 2 or more) kinds of information indicative of a transmission apparatus, a degree-of-similarity calculation unit that generates an i (an integer of 1 to K)-th sample feature from a radio feature, and calculates an i-th degree-of-similarity group, based on the i-th sample feature and an i-th set of the template feature group, a summed degree-of-similarity calculation unit that calculates a summed degree of similarity by summing K degrees of similarity by using an i-th weighting factor with respect to 1 to K of i, and an estimation unit that estimates that K information pieces, which are correlated in advance with calculation sources of K degrees of similarity having the summed degree of similarity that is highest, are information indicative of the transmission apparatus. 1. A transmission apparatus recognition apparatus comprising:a receiver configured to receive a signal wirelessly transmitted from a transmission apparatus;a radio feature generation unit configured to generate a radio feature from a received signal received by the receiver;a storage unit configured to store K sets of template feature groups for estimating K kinds of information indicative of the transmission apparatus, K being an integer of 2 or more;a degree-of-similarity calculation unit configured to generate an i-th sample feature from the radio feature, i being an integer of 1 to K, and configured to calculate an i-th degree-of-similarity group that is degrees of similarity between the i-th sample feature and template features included in an i-th set of the template feature group;a summed degree-of-similarity calculation unit configured to calculate a summed degree of similarity, by executing a process including a weighted-summing process of selecting, one by one, the degrees of similarity included in the i-th degree-of-similarity group, with ...

SYSTEMS AND METHODS FOR MEDICAL IMAGE SEGMENTATION AND ANALYSIS

The present disclosure includes systems, methods, and computer-readable medium for monitoring and analyzing skin lesions. A sequence of images are be received, and color correction, contour detection, and feature detection are performed on the images. A progression factor is determined based on a comparison of the an area of the lesion between images. A system for monitoring a progression of a skin lesion is provided that includes a portable imaging device to aid in capturing images of the lesion, and a user device configured to analyze the images and determine a progression factor of the skin lesion. 1. A system for measuring a body condition , the system comprising:an imaging sensor that detects images of a body region;a data processor that performs color correction on a first image of the body region, the data processor further performing histogram equalization on a color channel, that performs contour detection on the first image to identify one or more contours of the body region, and that performs feature detection on the first image to identify one or more features of the body region; anda memory device that stores image data.2. The system of claim 1 , wherein the image sensor is configured to detect a sequence of images wherein the images correspond to a skin lesion represented in the first image claim 1 , and the sequence of images represent the skin lesion over a period of time.3. The system of wherein the data processor performs color correction on the sequence of images including performing histogram equalization on a color channel; andthe data processor performs contour detection on the sequence of images to identify one or more contours of the skin lesion; andthe data processor performs feature detection on the sequence of images to identify one or more features of the skin lesion.4. (canceled)5. (canceled)6. The system of wherein the system stores the resulting sequence of images to record changes from a therapeutic procedure such as phototherapy.7. ...

Determining Data Representative of Bias Within a Model

Methods, systems, and computer program products for determining data representative of bias within a model are provided herein. A computer-implemented method includes obtaining a first dataset on which a model was trained, wherein the first dataset contains protected attributes, and a second dataset on which the model was trained, wherein the protected attributes have been removed from the second dataset; identifying, for each of the one or more protected attributes in the first dataset, one or more attributes in the second dataset correlated therewith; determining bias among at least a portion of the identified correlated attributes; and outputting, to at least one user, identifying information pertaining to the one or more instances of bias. 1. A computer-implemented method comprising:obtaining, as input, (i) a first dataset on which a model was trained, wherein the first dataset contains one or more protected attributes, and (ii) a second dataset on which the model was trained, wherein the one or more protected attributes have been removed from the second dataset;identifying, for each of the one or more protected attributes in the first dataset, one or more attributes in the second dataset correlated therewith; mapping at least two classes of data points associated with the correlated attributes in the second dataset to a set of values associated with the one or more protected attributes in the first dataset; and', 'identifying one or more instances of bias by observing a change to one or more of the values in the mappings in response to modifying one or more class designations among the data points in the mappings; and, 'determining bias among at least a portion of the identified correlated attributes, wherein said determining comprisesoutputting, to at least one user, identifying information pertaining to the one or more instances of bias;wherein the method is carried out by at least one computing device.2. The computer-implemented method of claim 1 , wherein ...

FETAL ELECTROCARDIOGRAPHIC SIGNAL PROCESSING METHOD AND FETAL ELECTROCARDIOGRAPHIC SIGNAL PROCESSING DEVICE

Provided are a fetal electrocardiographic signal processing method and a fetal electrocardiographic signal processing device that can appropriately select a signal reliably including a fetal electrocardiographic component from a plurality of signals separated by independent component analysis with reference. The fetal electrocardiographic signal processing method according to the invention includes: a fetal feature display signal extraction step of separating separation signals for a plurality of channels from biological signals of the plurality of channels acquired from a pregnant mother, using independent component analysis with reference, and removing noise from the separation signal for each channel to extract a fetal feature display signal; and a maternal electrocardiographic signal removal step of removing the fetal feature display signal at a timing when an electrocardiographic signal of the mother is likely to appear from the fetal feature display signals to obtain fetal feature signals including a large number of fetal electrocardiographic signals. 1. A fetal electrocardiographic signal processing method comprising:a fetal feature display signal extraction step of separating a separation signal for each channel from biological signals of a plurality of channels acquired from a pregnant mother, using independent component analysis with reference, and removing noise from each separation signal to extract a fetal feature display signal for each channel; anda maternal electrocardiographic signal removal step of removing the fetal feature display signal at a timing when an electrocardiographic signal of the mother is likely to appear from the fetal feature display signals to obtain fetal feature signals including a large number of fetal electrocardiographic signals.2. The fetal electrocardiographic signal processing method according to claim 1 , wherein the fetal feature display signal extraction step includes:a complex signal generation step of generating a ...

A Probability-Based Detector and Controller Apparatus, Method, Computer Program

An apparatus including circuitry configured to determine a probability by combining at least: a probability that an event is present within a current feature of interest given a first set of previous features of interest, and a probability that the event is present within the current feature of interest given a second set of previous features of interest, different to the first set of previous features of interest; circuitry configured to detect the event based on the determined probability; and circuitry configured to control, in dependence on the detection of the event, performance of an action.

IMAGE RECONSTRUCTION SYSTEM AND METHOD IN MAGNETIC RESONANCE IMAGING

A method and system for image reconstruction are provided. Multiple coil images may be obtained. A first reconstructed image based on the multiple coil images may be reconstructed based on a first reconstruction algorithm. A second reconstructed image based on the multiple coil images may be reconstructed based on a second reconstruction algorithm. Correction information about the first reconstructed image may be generated based on the first reconstructed image and the second reconstructed image. A third reconstructed image may be generated based on the first reconstructed image and the correction information about the first reconstructed image. 1. A method comprising:obtaining multiple coil images of an imaged object;reconstructing a first reconstructed image based on the multiple coil images according to a first reconstruction algorithm;reconstructing a second reconstructed image based on the multiple coil images according to a second reconstruction algorithm;generating correction information about the first reconstructed image based on the first reconstructed image and the second reconstructed image; andgenerating a third reconstructed image based on the first reconstructed image and the correction information about the first reconstructed image.2. The method of claim 1 , wherein the first reconstruction algorithm is a sum of squares algorithm.3. The method of claim 1 , wherein the second reconstruction algorithm is a geometric average algorithm.4. The method of claim 1 , wherein the reconstructing a first reconstructed image or the reconstructing a second reconstructed image further includes: determining pixel coordinates of corresponding pixels in the multiple coil images relating to the point of the imaged object; and', 'obtaining pixel values of the corresponding pixels in the multiple coil images of the point; and, 'for each point of a plurality of points in the imaged object'}reconstructing the first reconstructed image or the second reconstructed image ...

METHOD FOR DENOISING AN IMAGE AND APPARATUS FOR DENOISING AN IMAGE

Traditional image denoising requires image analysis or noise level analysis. Differently, the invention provides denoising by dividing an input image into small square overlapping patches, computing and storing the mean value of each patch, and subtracting it from the patch. Each zero-mean patch is then automatically aligned to a reference orientation by computing a few relevant DCT coefficients of the patch, analyzing the patch orientation in terms of transposition, inversion and horizontal and vertical flipping, and applying a re-orientation transform to automatically pose the patch in a standard orientation, regardless of its contents. The reoriented patches are clustered, and each of the resulting clusters is either shuffled or averaged. Then, all patches are re-transformed back to their original orientations by reversing the previous transforms, their respective mean is added and the denoised image is reconstructed by overlapping the re-transformed and mean added patches. 2. The method according to claim 1 , wherein said normalizing the patches refers to patch orientation and pixel values of a patch claim 1 , and comprises claim 1 , for a current patch claim 1 , at least one ofmean subtraction, wherein a mean value of the pixel values of the current patch is subtracted from said pixel values;orientation normalization, wherein a spatial orientation of the mean-subtracted current patch is determined, and wherein the mean-subtracted current patch is transposed, rotated and/or flipped to a normal orientation; and{'b': '33', 'inversion, wherein the pixel values of said mean-subtracted current patch are inverted, before or after or during said orientation normalization ().'}3. The method according to claim 2 , further comprising calculating DCT coefficients of said mean-subtracted current patch claim 2 , wherein the DCT coefficients indicate whether the mean-subtracted current patch is to be inverted claim 2 , transposed claim 2 , rotated and/or flipped for obtaining ...

SYSTEMS, METHODS, AND DEVICES FOR MEDICAL IMAGE ANALYSIS, DIAGNOSIS, RISK STRATIFICATION, DECISION MAKING AND/OR DISEASE TRACKING

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters. 130.-. (canceled)31. A computer-implemented method for assessing risk of coronary artery disease of a subject based on image analysis of a non-invasive medical image of the subject and compound analysis of vascular parameters and plaque parameters derived from the image analysis , the method comprising:accessing, by a computer system, a medical image of a coronary region of a subject, wherein the medical image of the coronary region of the subject is obtained non-invasively;identifying, by the computer system, one or more coronary arteries and one or more regions of plaque in the medical image of the coronary region of the subject;determining, by the computer system, one or more vascular morphology parameters and one or more quantified plaque parameters of the one or more identified regions of plaque, wherein the one or more quantified plaque parameters comprises a ratio or function of volume or surface area of plaque and composition of plaque, the composition of plaque derived based on analysis of radiodensity values corresponding to the one or more identified regions of plaque, wherein the composition of plaque comprises one or more of non-calcified ...

FACILITATING AUTOMATIC HANDLING OF INCOMPLETE DATA IN A RANDOM FOREST MODEL

Techniques are provided for training and/or executing, by a system operatively coupled to a processor, a modified random forest model using a process that employs significance of data fields in performing imputation, filtering data records out of sample datasets for generating subtrees, and filtering out subtrees for making predictions. 1. A system , comprising:a memory that stores computer executable components; [ determines whether data fields of a dataset are deemed to be significant based on a significance function,', 'labels a first set of the data fields that are determined to be significant with an indication of being a significant data field, and', 'labels a second set of the data fields that are determined not to be significant with an indication of being a non-significant data field; and, 'a significance component that, 'a training component that trains a modified random forest model based on a training process that employs the indication of being the significant data field and the indication of being the non-significant data field., 'a processor, operably coupled to the memory, and that executes computer executable components stored in the memory, wherein the computer executable components comprise2. The system of claim 1 , wherein the computer executable components further comprise an imputation component that imputes claim 1 , during the training process claim 1 , data values for ones of the second set of the data fields and that are missing data values in data records of the dataset.3. The system of claim 2 , wherein the computer executable components further comprise a sampling component that generates sample datasets from the dataset with respective sample data fields from the data fields.4. The system of claim 3 , wherein the sampling component further:filters out, during the training process, from a sample dataset of the sample datasets, a data record having a data field from the first set and the data field is missing a data value.5. The system of ...

CHART FOR EVALUATING SKIN COLOR AND ITS APPLICATION TO EFFICACY EVALUATION OF ANTI-AGING AND SKIN LIGHTENING PRODUCTS

The present invention relates to charts, stacks and methods for evaluating skin color of a mammalian subject. The chart may include a substrate and an indicia visible from a first side of the substrate that includes a plurality of images of mammalian skin tones having varying degrees of yellowness, wherein each indicia is correlated with an index value. Also, the present invention relates to packaged topical cosmetic products that include a skin chart as well as to methods for evaluating anti-aging and skin lightening products. 114.-. (canceled)15. A method of creating a chart comprising:(i) analyzing a plurality of facial images for their color properties: skin brightness (L*), redness (a*), and yellowness (b*);(ii) measuring the L*, a* and b* values for each analyzed facial image;(iii) sorting the measured b* values to show a low-to-high scale;(iv) normalizing the measured b* values to the +/−4 sigma range to form a skin yellowness distribution of 1-10 point scale range;(v) selecting target values for the L* and a* values;(vi) selecting facial images that correspond to the skin yellowness distribution of 1-10 point scale range, wherein the a* and the L* values are target values for the different b* values at each scale point;(vii) printing representative facial images corresponding to the skin yellowness distribution of 1-10 point scale onto a substrate.16. The method of claim 15 , wherein the target value for L* is 69 and the target value for a* is 14.8 when choosing various scale points for the b* values.17. The method of claim 15 , wherein the chart is for evaluating of anti-aging and skin lightning products by comparing a skin tone of a mammalian subject to the chart.18. The method of claim 15 , wherein the chart is for evaluating skin color of a mammalian subject.19. The method of claim 15 , wherein the chart is for evaluating skin yellowness of a mammalian subject.20. The method of claim 15 , wherein the step (vi) comprises selecting the images as follows:a ...

Method for training cross-modal retrieval model, electronic device and storage medium

The present disclosure discloses a method for training a cross-modal retrieval model, an electronic device and a storage medium, and relates to the field of computer technologies, and particularly to the field of artificial intelligence technologies, such as knowledge graph technologies, computer vision technologies, deep learning technologies, or the like. The method for training a cross-modal retrieval model includes: determining similarity of a cross-modal sample pair according to the cross-modal sample pair, the cross-modal sample pair including a sample of a first modal and a sample of a second modal, and the first modal being different from the second modal; determining a soft margin based on the similarity, and determining a soft margin loss function based on the soft margin; and determining a total loss function based on the soft margin loss function, and training a cross-modal retrieval model according to the total loss function.

DEVICE FOR OPTIMIZING TRAINING INDICATOR OF ENVIRONMENT PREDICTION MODEL, AND METHOD FOR OPERATING SAME

The present invention relates to an apparatus for optimizing training indicators of an environmental prediction model and an operation method thereof. A training indicator optimization apparatus according to an embodiment includes a pre-processor for constructing a base dataset for environmental measurement data; a dynamic feature processor for identifying and extracting dynamic features for the constructed base dataset through multi-resolution wavelet analysis and a dimensionality reduction technique; a key feature group selector for identifying and evaluating driving force for environmental measurement data based on the extracted dynamic features and selecting a key feature group in response to the evaluation result; and an indicator optimizer for receiving the selected key feature group and the environmental measurement data as inputs and controlling a plurality of training indicators corresponding to an environmental prediction model. 1. An apparatus for optimizing training indicators , comprising:a pre-processor for constructing a base dataset for environmental measurement data;a dynamic feature processor for identifying and extracting dynamic features for the constructed base dataset through multi-resolution wavelet analysis and a dimensionality reduction technique;a key feature group selector for identifying and evaluating driving force for environmental measurement data based on the extracted dynamic features and selecting a key feature group in response to the evaluation result; andan indicator optimizer for receiving the selected key feature group and the environmental measurement data as inputs and controlling a plurality of training indicators corresponding to an environmental prediction model.2. The apparatus according to claim 1 , wherein the environmental measurement data comprises hydrological-environmental time series data measured in real time claim 1 , the hydrological-environmental time series data comprising at least one environmental data of ...

SYSTEMS AND METHODS FOR DETERMINING REFERENCE POINTS FOR MACHINE LEARNING ARCHITECTURES

This disclosure relates to improved techniques for determining reference points for computerized simulations of physical systems and/or physical models that may be used in machine learning development architectures. This disclosure also relates to systems, methods, apparatuses, and computer program products that are configured to determine reference points for one or more parameters of a model of a physical system used in a computerized simulation of the model. The reference points may be representative of the system outputs across the parameter space, and can be determined in an efficient and computationally-feasible manner. The outputs of the computerized simulations of physical systems may then be further used to create, build, or train one or more learning models pertaining to physical systems. 1. A system of one or more computing devices comprising one or more processors and one or more non-transitory storage devices for storing instructions , wherein execution of the instructions by the one or more processors causes the one or more computing devices to:receive a parameter space definition;determine a correlation metric on a parameter space using the parameter space definition;determine a loss function using the correlation metric;compute a set of reference points using the loss function;generate one or more sensed outputs using the computed reference points; andupdate a learning model of a machine learning development architecture using a training vector comprised of the reference points and sensed outputs.2. The system of claim 1 , wherein the correlation metric is determined using a derivative of a scalar function claim 1 , and the scalar function assigns a scalar value to every point in the parameter space to approximate one or more of the sensed outputs.3. The system of claim 2 , wherein the scalar function is a geometric function of the points in the parameter space claim 2 , and the geometric function includes a volume or an area.4. The system of claim 2 ...

SYSTEMS, METHODS, AND DEVICES FOR MEDICAL IMAGE ANALYSIS, DIAGNOSIS, RISK STRATIFICATION, DECISION MAKING AND/OR DISEASE TRACKING

The disclosure herein relates to systems, methods, and devices for medical image analysis, diagnosis, risk stratification, decision making and/or disease tracking. In some embodiments, the systems, devices, and methods described herein are configured to analyze non-invasive medical images of a subject to automatically and/or dynamically identify one or more features, such as plaque and vessels, and/or derive one or more quantified plaque parameters, such as radiodensity, radiodensity composition, volume, radiodensity heterogeneity, geometry, location, and/or the like. In some embodiments, the systems, devices, and methods described herein are further configured to generate one or more assessments of plaque-based diseases from raw medical images using one or more of the identified features and/or quantified parameters. 130.-. (canceled)31. A computer-implemented method for displaying computed tomography (CT) images and corresponding coronary vessel information including images rendered from the CT images and identification of lumen , vessel walls , and plaque of coronary vessels determined from the CT images by image processing , the method comprising:accessing, by a computer system, a set of CT images of coronary vessels of a subject and coronary vessel information associated with the set of CT images, the coronary vessel information including identification of lumen, vessel walls, and plaque of one or more coronary vessels;generating and displaying, by the computer system, in a user interface a first panel illustrating at least a portion of a coronary vessel of the subject in a plurality of straightened multiplanar (SMPR) vessel views comprising the coronary vessel information, the plurality of SMPR vessel views adjacently positioned to one another in the first panel, each of the plurality of SMPR vessel views rotationally offset from one another by a predetermined rotational angle along a longitudinal axis of the plurality of SMPR views;generating and displaying, ...

SYSTEMS AND METHODS FOR X-RAY IMAGE PASTING

A method includes receiving a first image and a second image from an X-ray imaging device and determining a template window in the first image and a plurality of search windows in the second image. The method further includes generating a template vector corresponding to the template window and a plurality of search vectors corresponding to the plurality of search windows. The method also includes calculating a plurality of similarity scores based on the template vector and the plurality of search vectors. Additionally, the method includes determining a matching window from the plurality of search windows based on the plurality of similarity scores. Finally, the method includes generating a final image using the first image and the second image based on the template window and the matching window. 1. A method , comprising:receiving a first image and a second image from an X-ray imaging device;determining a template window in the first image and a plurality of search windows in the second image;generating a template vector corresponding to the template window and a plurality of search vectors corresponding to the plurality of search windows;calculating a plurality of similarity scores based on the template vector and the plurality of search vectors;determining a matching window from the plurality of search windows based on the plurality of similarity scores; andgenerating a final image using the first image and the second image based on the template window and the matching window.2. The method of claim 1 , further comprising:extracting a first X-ray dose associated with the first image and a second X-ray dose associated with the second image;calculating a dose ratio based on the first X-ray dose and the second X-ray dose; andnormalizing the first image based on the dose ratio.3. The method of claim 1 , further comprising:extracting a first detector position associated with the first image and a second detector position associated with to the second image; ...

SELF-TUNING EVENT DETECTION

A method is provided for processing environmental sensor data comprising the following steps. One or more raw data values are received from an environmental sensor, an average value and a measure of dispersion are determined over a defined time period only from raw data values between a lower threshold and an upper threshold, and the lower threshold and the upper threshold are redefined depending on the average value and the measure of dispersion. The method may be used in an environmental sensor, e.g. a MOX sensor or a VOC sensor, and implemented as a computer program. 1. Method for processing environmental sensor data , comprising the steps ofreceiving from an environmental sensor one or more raw data values,determining an average value and a measure of dispersion over a defined time period only from raw data values between a lower threshold and an upper threshold,redefining the lower threshold and the upper threshold depending on the average value and the measure of dispersion,in response to a gating adaptation event determining the average value and the measure of dispersion dependent on the raw data value received for the corresponding time step even if not between the lower threshold and the upper threshold.2. Method according to claim 1 ,wherein one of the raw data values is received per discrete time step, wherein the average value and the measure of dispersion are determined anew per time step dependent on the raw data value received at the corresponding time step, wherein the lower threshold and the upper threshold are redefined anew per time step dependent on the average value and the measure of dispersion determined for the corresponding time step,3. Method according to claim 2 ,wherein the lower threshold is the average value minus the measure of dispersion,wherein the upper threshold is the average value plus the measure of dispersion,wherein the measure of dispersion corresponds to one or two times a standard deviation of the raw data values, and/or ...

ACCURATE ROI EXTRACTION AIDED BY OBJECT TRACKING

An image data processing method includes receiving frame image data of N frames, where N>1, detecting a region of interest in one of the N frames, tracking locations of the region of interest in at least one of the N frames, and providing a merged location of the region of interest based on the locations of the region of interest in the N frames. Some embodiments include providing T of the merged locations of the region of interest for T respective groups of N frames, where T>1, providing respective statistical data for each of the T merged locations, and providing a final location of the region of interest based on the T merged locations and the statistical data for the T merged locations. 1. An image data processing method comprising:receiving frame image data of N frames, where N>1;detecting a region of interest in one of the N frames;tracking locations of the region of interest in at least one of the N frames; andproviding a merged location of the region of interest based on the locations of the region of interest in the N frames.2. The image data processing method of claim 1 , further comprising:providing T of the merged locations of the region of interest for T respective groups of N frames, where T>1;providing respective statistical data for each of the T merged locations; andproviding a final location of the region of interest based on the T merged locations and the statistical data for the T merged locations.3. The image data processing method of claim 2 , wherein the statistical data for each of the T merged locations comprises at least one of:a number of the frames in which the region of interest appeared; anda percentage of the frames in which the region of interest appeared.4. The image data processing method of claim 1 , further comprising:receiving a previously-detected location for the region of interest; andproviding the final location of the region of interest based on the previously-detected location, the T merged locations, and the statistical ...