weixin_46179086的博客

The video-based HOI anticipation task in the third-person view is rarely researched. In this paper, a framework to detect current HOIs and anticipate future HOIs in videos is propose.

To address the case for human-object interactions in-the-wild videos with high degree of label skew, SERVO-HOI, a robust end-to-end framework is proposed. This network contextualizes multiple image representations and is trained to explicitly handle datase

This work propose the multiplex relation network (MUREN) that performs rich context exchange between three decoder branches using unary, pairwise, and ternary relations of human, object, and interaction tokens.