Target tracking method and tracker based on decoupling classification and regression features

The invention discloses a target tracking method and tracker based on decoupling classification and regression features. The tracking method comprises the following steps: dividing a target template region and a target search region according to an initial frame and a current frame in a video sequence; deep learning characterization features of the target template region and the target search region are extracted by using a twin deep residual convolutional neural network; fusing the target template region depth representation feature and the target search region depth representation feature based on a transform cross attention mechanism and convolution operation to obtain a decoupling classification feature and a decoupling regression feature of a tracker; the decoupling classification features and the decoupling regression features are further coded based on a transform self-attention mechanism and convolution operation, and target foreground and background dichotomy prediction and target ...