分离静态与动态信息，缓解动作识别中的静态偏见问题。
Disentangling Static and Dynamic Information for Reducing Static Bias in Action Recognition
- 通过独立性损失分离动态与静态信息流
- 显著降低模型对静态场景的依赖性
- 适合提升真实场景和零样本动作识别性能
动作识别模型过度依赖静态线索而非动态人体运动，这种现象称为静态偏见，导致实际应用和零样本识别性能下降。本文提出一种方法，通过将时间动态信息与静态场景信息分离来缓解静态偏见。该方法结合了有偏与无偏信息流之间的统计独立性损失以及场景预测损失。实验表明，该方法有效减少了静态偏见，并验证了场景预测损失的重要性。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。