Yang Weidong

TL;DR: In this paper, a face detection method and system based on the flexible convolution network is proposed, which consists of steps that an input image is received, and the image includes the to-be-detected face information; features of the image are extracted through the main network, and features include semantic features and offset features; the semantic features, the offset features and the features are respectively outputted through flexible convolutions.

TL;DR: In this paper, a target tracking method and system of a convolution twin network based on multi-layer feature fusion was proposed, which integrates scores of different layers, combines high-level semantic features and bottom-level detail features, better distinguishes the interference of similar or similar targets, and prevents the problems of target drift and target loss in the tracking process.

TL;DR: In this article, a multi-scale label-based sub-pixel convolution image super-resolution reconstruction method is proposed, which consists of a feature extraction structure, a residual error module, an up-sampling structure and a cross-scale jump connection structure.

TL;DR: In this article, a target tracking method based on a fusion of a correlation filtering and a twin convolutional network is proposed, which comprises the following steps of: extracting a target feature map of the (t-1) frame image known target position by using the first convolution network; extracting a search feature maps of the t-1 frame image using the second convolution neural network; obtaining a target region of the 2D frame image by performing a fast Fourier transform on the target feature maps; and calculating the cross-correlation between the search region of t-