| Title |
Improving Action Recognition via Specialized Self-supervised methods Collaboration |
| Authors |
김태훈(Taehoon Kim) ; 황원준(wonj니n Hwang) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.2.101 |
| Keywords |
Computer vision; Pattern recognition; Action recognition; Self-supervised learning |
| Abstract |
In this paper, we propose a method to simultaneously learn two self-supervised Pretext tasks to enhance the performance of Action Recognition. Pretext tasks should be specialized independently in spatial and temporal part, and the method is proposed to be designed to affect each other in the learning process so that deep learning models can learn more abundant representation at the same time. So we conducted an experiment to verify the validity of the proposed training procedure(strategy). The experiment was conducted in three ways: Accuracy comparison with previous methods, Video Retrieval, and Ablation study for a spatial pretext task effective for video. |