Temporal action localization aims to localize starting and ending time with action category. Limited by GPU memory, mainstream methods pre-extract features for each video. Therefore, feature quality determines the upper bound of detection performance. In this technical report, we explored classic convolution-based backbones and the recent surge of transformer-based backbones. We found that the transformer-based methods can achieve better classification performance than convolution-based, but they cannot generate accuracy action proposals. In addition, extracting features with larger frame resolution to reduce the loss of spatial information can also effectively improve the performance of temporal action localization. Finally, we achieve 42.42% in terms of mAP on validation set with a single SlowFast feature by a simple combination: BMN+TCANet, which is 1.87% higher than the result of 2020's multi-model ensemble. Finally, we achieve Rank 1st on the CVPR2021 HACS supervised Temporal Action Localization Challenge.
@article{arxiv.2106.13014,
title = {Exploring Stronger Feature for Temporal Action Localization},
author = {Zhiwu Qing and Xiang Wang and Ziyuan Huang and Yutong Feng and Shiwei Zhang and jianwen Jiang and Mingqian Tang and Changxin Gao and Nong Sang},
journal= {arXiv preprint arXiv:2106.13014},
year = {2021}
}
Comments
Rank 1st on the CVPR2021 HACS supervised Temporal Action Localization Challenge