Generative methods of super resolution
Номер патента: EP3298576A1
Опубликовано: 28-03-2018
Автор(ы): Ferenc Huszar, Rob BISHOP, Wenzhe Shi, Zehan WANG
Принадлежит: Magic Pony Technology Ltd
Опубликовано: 28-03-2018
Автор(ы): Ferenc Huszar, Rob BISHOP, Wenzhe Shi, Zehan WANG
Принадлежит: Magic Pony Technology Ltd
Реферат: A method for training an algorithm to process at least a section of received visual data using a training dataset and reference dataset. The method comprises an iterative method with each iteration comprising the steps of: generating a set of training data using the algorithm; comparing one or more characteristics of the training data to one or more characteristics of at least a section of the reference dataset; and modifying one or more parameters of the algorithm to optimise processed visual data based on the comparison between the characteristic of the training data and the characteristic of the reference dataset. The algorithm may output the processed visual data with the same content as the at least a section of received visual data. Some aspects and/or embodiments provide for improved super-resolution of lower quality images, with a view to producing super-resolution images which have improved characteristics (e.g. less blur, less undesired smoothing) compared to other super-resolution techniques.
Super-resolution using time-space-frequency tokens
Номер патента: WO2023240609A1. Автор: Huan Yang,Jianlong FU. Владелец: Microsoft Technology Licensing, LLC. Дата публикации: 2023-12-21.