Action recognition models rely excessively on static cues rather than dynamic human motion, which is known as static bias. This bias leads to poor performance in real-world applications and zero-shot action recognition. In this paper, we propose a method to reduce static bias by separating temporal dynamic information from static scene information. Our approach uses a statistical independence loss between biased and unbiased streams, combined with a scene prediction loss. Our experiments demonstrate that this method effectively reduces static bias and confirm the importance of scene prediction loss.
@article{arxiv.2509.23009,
title = {Disentangling Static and Dynamic Information for Reducing Static Bias in Action Recognition},
author = {Masato Kobayashi and Ning Ding and Toru Tamaki},
journal= {arXiv preprint arXiv:2509.23009},
year = {2025}
}
Comments
in Proc. of ICCV2025 Workshop and Challenge on Disentangled Representation Learning for Controllable Generation (DRL4Real)