Research map / Label-efficient and robust learning / Recognition and long-tail
Learning View-Disentangled Human Pose Representation by Contrastive Cross-View Mutual Information Maximization
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use CV-MIM instead of Pr-VIPE | Penn Action single-shot cross-view action recognition · Penn Action single-shot cross-view | average top-1 accuracy (%): 89.03 ± 0.40 -> 91.75 ± 0.24 | Helped |
| Use CV-MIM instead of Pr-VIPE | NTU-RGB+D single-shot cross-view action recognition · NTU-RGB+D single-shot cross-view | average top-1 accuracy (%): 54.8 -> 56.6 | Helped |
| Use CV-MIM instead of the strongest listed pose baseline | Penn Action fully supervised action recognition · Penn Action fully supervised | top-1 accuracy (%): 98.8 -> 98.1 | Hurt |
| Use CV-MIM instead of ST-GCN | NTU-RGB+D cross-view action recognition · NTU-RGB+D cross-view | top-1 accuracy (%): 91.3 -> 89.5 | Hurt |
| Replace concatenation fusion with mixture-of-experts fusion (ablation) | CV-MIM · Penn Action single-shot cross-view | accuracy (%): 90.5 -> 91.8 | Helped |
| Replace concatenation fusion with product fusion (ablation) | CV-MIM · Penn Action single-shot cross-view | accuracy (%): 90.5 -> 90.2 | Hurt |
| Add the disentanglement loss (ablation) | CV-MIM · Penn Action single-shot cross-view | accuracy (%): 86.1 -> 91.8 | Helped |
| Add the prior-matching loss (ablation) | CV-MIM · Penn Action single-shot cross-view | accuracy (%): 89.3 -> 91.8 | Helped |
| Add camera augmentation (ablation) | CV-MIM · Penn Action single-shot cross-view | accuracy (%): 90.75 ± 0.30 -> 91.75 ± 0.24 | Helped |
| Add camera augmentation (ablation) | CV-MIM · NTU-RGB+D single-shot cross-view | accuracy (%): 54.50 ± 0.22 -> 56.50 ± 0.13 | Helped |
Closest papers by meaning
- DVANet: Disentangling View and Action Features for Multi-View Action Recognition (AAAI 2024 · similarity 0.87)
- PeVL: Pose-Enhanced Vision-Language Model for Fine-Grained Human Action Recognition (CVPR 2024 · similarity 0.87)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.86)
- Dual Pose-invariant Embeddings: Learning Category and Object-specific Discriminative Representations for Recognition and Retrieval (CVPR 2024 · similarity 0.85)
- SVTformer: Spatial-View-Temporal Transformer for Multi-View 3D Human Pose Estimation (AAAI 2025 · similarity 0.85)
- Cross-view and Cross-pose Completion for 3D Human Understanding (CVPR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.