Action Coherence Network for Weakly Supervised Temporal Action Localization

Yuanhao Zhai, Le Wang, Ziyi Liu, Qilin Zhang, Gang Hua, Nanning Zheng

Abstract

Most prominent temporal action localization methods are of the fully-supervised type, which rely heavily on frame-level labels, which could be prohibitively expensive to annotate. Thanks to recent developments on the Weakly-supervised Temporal Action Localization (W-TAL), this alternative paradigm requires only video-level labels in training, alleviating such annotation efforts. Specifically, we present Action Coherence Network (ACN) for W-TAL, which features a new coherence loss that better supervises action boundary learning and facilitate proposal regression. In addition, a purpose-built fusion module is proposed for localization inference based on features extracted by two streams of convolutional neural network. Overall, the proposed ACN achieves state-of-the-art W-TAL performance on two challenging datasets (THU-MOS14 and ActivityNet1.2, particularly ACN attains mAP of 24.2% on THUMOS14 under IoU threshold 0.5), which is approaching some recent fully-supervised TAL methods.
BibTeX
@inproceedings{zhai2019action,
  author = {Yuanhao Zhai and Le Wang and Ziyi Liu and Qilin Zhang and Gang Hua and Nanning Zheng},
  title = {{Action Coherence Network for Weakly Supervised Temporal Action Localization}},
  booktitle = {ICIP},
  year = {2019},
  doi = {10.1109/icip.2019.8803447},
}

Loading related content...

Figure