Relation Modeling in Spatio-Temporal Action Localization

15 Jun 2021  ·  Yutong Feng, Jianwen Jiang, Ziyuan Huang, Zhiwu Qing, Xiang Wang, Shiwei Zhang, Mingqian Tang, Yue Gao ·

This paper presents our solution to the AVA-Kinetics Crossover Challenge of ActivityNet workshop at CVPR 2021. Our solution utilizes multiple types of relation modeling methods for spatio-temporal action detection and adopts a training strategy to integrate multiple relation modeling in end-to-end training over the two large-scale video datasets. Learning with memory bank and finetuning for long-tailed distribution are also investigated to further improve the performance. In this paper, we detail the implementations of our solution and provide experiments results and corresponding discussions. We finally achieve 40.67 mAP on the test set of AVA-Kinetics.

PDF Abstract

Code
Add Remove Mark official

No code implementations yet. Submit your code now

Datasets

Kinetics AVA

Results from the Paper
Edit

Ranked #4 on Spatio-Temporal Action Localization on AVA-Kinetics (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data 		Result Benchmark
Spatio-Temporal Action Localization AVA-Kinetics RM (multi-scale, ensemble) val mAP 40.52 # 4
Compare
Spatio-Temporal Action Localization AVA-Kinetics RM (multi-scale, ir-CSN-152) val mAP 37.95 # 6
Compare

Methods
Add Remove

No methods listed for this paper. Add relevant methods here
Contact us on: hello@paperswithcode.com . Papers With Code is a free resource with all data licensed under CC-BY-SA.
Terms Data policy Cookies policy