Hierarchical Attention Network for Action Recognition in Videos
This addresses the challenge of understanding complex human actions in wild videos, which has broad applications, but it appears incremental as it builds on existing CNN-based approaches with attention mechanisms.
The paper tackles the problem of human action recognition in videos by proposing a Hierarchical Attention Network (HAN) that incorporates spatial, short-term motion, and long-term temporal structures, achieving state-of-the-art performance on UCF-101 and HMDB-51 benchmarks.
Understanding human actions in wild videos is an important task with a broad range of applications. In this paper we propose a novel approach named Hierarchical Attention Network (HAN), which enables to incorporate static spatial information, short-term motion information and long-term video temporal structures for complex human action understanding. Compared to recent convolutional neural network based approaches, HAN has following advantages (1) HAN can efficiently capture video temporal structures in a longer range; (2) HAN is able to reveal temporal transitions between frame chunks with different time steps, i.e. it explicitly models the temporal transitions between frames as well as video segments and (3) with a multiple step spatial temporal attention mechanism, HAN automatically learns important regions in video frames and temporal segments in the video. The proposed model is trained and evaluated on the standard video action benchmarks, i.e., UCF-101 and HMDB-51, and it significantly outperforms the state-of-the arts