MAGR: Manifold-Aligned Graph Regularization for Continual Action Quality Assessment
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916465302568960 |
|---|---|
| author | Zhou, Kanglei Wang, Liyuan Zhang, Xingxing Shum, Hubert P. H. Li, Frederick W. B. Li, Jianguo Liang, Xiaohui |
| author_facet | Zhou, Kanglei Wang, Liyuan Zhang, Xingxing Shum, Hubert P. H. Li, Frederick W. B. Li, Jianguo Liang, Xiaohui |
| contents | Action Quality Assessment (AQA) evaluates diverse skills but models struggle with non-stationary data. We propose Continual AQA (CAQA) to refine models using sparse new data. Feature replay preserves memory without storing raw inputs. However, the misalignment between static old features and the dynamically changing feature manifold causes severe catastrophic forgetting. To address this novel problem, we propose Manifold-Aligned Graph Regularization (MAGR), which first aligns deviated old features to the current feature manifold, ensuring representation consistency. It then constructs a graph jointly arranging old and new features aligned with quality scores. Experiments show MAGR outperforms recent strong baselines with up to 6.56%, 5.66%, 15.64%, and 9.05% correlation gains on the MTL-AQA, FineDiving, UNLV-Dive, and JDM-MSA split datasets, respectively. This validates MAGR for continual assessment challenges arising from non-stationary skill variations. Code is available at https://github.com/ZhouKanglei/MAGR_CAQA}{https://github.com/ZhouKanglei/MAGR_CAQA. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_04398 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MAGR: Manifold-Aligned Graph Regularization for Continual Action Quality Assessment Zhou, Kanglei Wang, Liyuan Zhang, Xingxing Shum, Hubert P. H. Li, Frederick W. B. Li, Jianguo Liang, Xiaohui Computer Vision and Pattern Recognition Action Quality Assessment (AQA) evaluates diverse skills but models struggle with non-stationary data. We propose Continual AQA (CAQA) to refine models using sparse new data. Feature replay preserves memory without storing raw inputs. However, the misalignment between static old features and the dynamically changing feature manifold causes severe catastrophic forgetting. To address this novel problem, we propose Manifold-Aligned Graph Regularization (MAGR), which first aligns deviated old features to the current feature manifold, ensuring representation consistency. It then constructs a graph jointly arranging old and new features aligned with quality scores. Experiments show MAGR outperforms recent strong baselines with up to 6.56%, 5.66%, 15.64%, and 9.05% correlation gains on the MTL-AQA, FineDiving, UNLV-Dive, and JDM-MSA split datasets, respectively. This validates MAGR for continual assessment challenges arising from non-stationary skill variations. Code is available at https://github.com/ZhouKanglei/MAGR_CAQA}{https://github.com/ZhouKanglei/MAGR_CAQA. |
| title | MAGR: Manifold-Aligned Graph Regularization for Continual Action Quality Assessment |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2403.04398 |