MFI: Multi-range feature interchange for video action recognition

Sikai Bai, Qi Wang, Xuelong Li

科研成果: 书/报告/会议事项章节会议稿件同行评审

4 引用 (Scopus)

摘要

Short-range motion features and long-range dependencies are two complementary and vital cues for action recognition in videos, but it remains unclear how to efficiently and effectively extract these two features. In this paper, we propose a novel network to capture these two features in a unified 2D framework. Specifically, we first construct a Short-range Temporal Interchange (STI) block, which contains a Channels-wise Temporal Interchange (CTI) module for encoding short-range motion features. Then a Graph-based Regional Interchange (GRI) module is built to present long-range dependencies using graph convolution. Finally, we replace original bottleneck blocks in the ResNet with STI blocks and insert several GRI modules between STI blocks, to form a Multi-range Feature Interchange (MFI) Network. Practically, extensive experiments are conducted on three action recognition datasets (i.e., Something-Something V1, HMDB51, and UCF101), which demonstrate that the proposed MFI network achieves impressive results with very limited computing cost.

源语言英语
主期刊名Proceedings of ICPR 2020 - 25th International Conference on Pattern Recognition
出版商Institute of Electrical and Electronics Engineers Inc.
6664-6671
页数8
ISBN(电子版)9781728188089
DOI
出版状态已出版 - 2020
活动25th International Conference on Pattern Recognition, ICPR 2020 - Virtual, Milan, 意大利
期限: 10 1月 202115 1月 2021

出版系列

姓名Proceedings - International Conference on Pattern Recognition
ISSN(印刷版)1051-4651

会议

会议25th International Conference on Pattern Recognition, ICPR 2020
国家/地区意大利
Virtual, Milan
时期10/01/2115/01/21

指纹

探究 'MFI: Multi-range feature interchange for video action recognition' 的科研主题。它们共同构成独一无二的指纹。

引用此