Jingyi Yu

TL;DR: Zhang et al. as discussed by the authors proposed a multi-stream spatial-temporal graph convolutional network (ST-GCN) to predict the fine-grained semantic action attributes, and adopted a semantic attribute mapping block to assemble various correlated action attributes into a high-level action label for the overall detailed understanding of the whole sequence, so as to enable various applications like action assessment or motion scoring.

TL;DR: A generic multiview tracking (GMT) framework that allows camera movement, while requiring neither specific object model nor camera calibration, is proposed, which addresses missing target issues such as occlusion.

TL;DR: In this article, an end-to-end deep learning scheme is proposed to establish dense shape correspondences and subsequently compress the data, which achieves state-of-the-art performance on both public and newly captured datasets.

TL;DR: This work combines deep learning and stereo matching to provide each ray a semantic label and design tailored weighting schemes for blending the rays, which can effectively remove foreground residues when focusing on the background.

TL;DR: A novel stereo trifocal tensor solver is presented and the camera matrix’s ability to continuously and robustly bootstrap visual motion estimation pipelines via integration into a robust, purely line-based visual odometry pipeline is outlined.