no code implementations • CVPR 2023 • Zhao Xie, Tian Gao, Kewei Wu, Jiao Chang
Third, we describe the nodes (actors) with causality features and learn the edges by fusing the causality relation with the appearance relation and distance relation.
no code implementations • TMM 2022 • Zhao Xie, Jiansong Chen, Kewei Wu, Dan Guo, Richang Hong
In the global aggregation module, the global prior knowledge is learned by aggregating the visual feature sequence of video into a global vector.
Ranked #62 on Action Recognition on Something-Something V2