Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

TEINet: Towards an Efficient Architecture for Video Recognition

About

Efficiency is an important issue in designing video architectures for action recognition. 3D CNNs have witnessed remarkable progress in action recognition from videos. However, compared with their 2D counterparts, 3D convolutions often introduce a large amount of parameters and cause high computational cost. To relieve this problem, we propose an efficient temporal module, termed as Temporal Enhancement-and-Interaction (TEI Module), which could be plugged into the existing 2D CNNs (denoted by TEINet). The TEI module presents a different paradigm to learn temporal features by decoupling the modeling of channel correlation and temporal interaction. First, it contains a Motion Enhanced Module (MEM) which is to enhance the motion-related features while suppress irrelevant information (e.g., background). Then, it introduces a Temporal Interaction Module (TIM) which supplements the temporal contextual information in a channel-wise manner. This two-stage modeling scheme is not only able to capture temporal structure flexibly and effectively, but also efficient for model inference. We conduct extensive experiments to verify the effectiveness of TEINet on several benchmarks (e.g., Something-Something V1&V2, Kinetics, UCF101 and HMDB51). Our proposed TEINet can achieve a good recognition accuracy on these datasets but still preserve a high efficiency.

Zhaoyang Liu, Donghao Luo, Yabiao Wang, Limin Wang, Ying Tai, Chengjie Wang, Jilin Li, Feiyue Huang, Tong Lu• 2019

Related benchmarks

TaskDatasetResultRank
Action RecognitionSomething-Something v2 (val)
Top-1 Accuracy66.5
535
Action RecognitionKinetics-400
Top-1 Acc76.2
413
Action RecognitionUCF101
Accuracy96.7
365
Action RecognitionSomething-Something v2
Top-1 Accuracy66.5
341
Action RecognitionSomething-Something v2 (test)
Top-1 Acc66.6
333
Action RecognitionSomething-something v1 (val)
Top-1 Acc52.5
257
Action RecognitionKinetics 400 (test)
Top-1 Accuracy76.2
245
Video ClassificationKinetics 400 (val)
Top-1 Acc76.2
204
Action RecognitionSomething-something v1 (test)
Top-1 Accuracy52.5
189
Action RecognitionSomething-Something v2 (test val)
Top-1 Accuracy65.5
187
Showing 10 of 26 rows

Other info

Follow for update