Research map / Video / Action recognition
Unsupervised Video Domain Adaptation for Action Recognition: A Disentanglement Perspective
Read the paperCheck 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 |
|---|---|---|---|
| Disentangle static domain-specific factors from dynamic action-related factors, with temporal adversarial alignment | TranSVAE · UCF-HMDB average | accuracy: 91.79 -> 93.37 | Helped |
| Add task-specific classification supervision (ablation) | TranSVAE · UCF-HMDB average | accuracy: 28.83 -> 63.19 | Helped |
| Add adversarial temporal alignment (ablation) | TranSVAE · UCF-HMDB average | accuracy: 55.38 -> 63.19 | Helped |
| Add mutual dependence minimization (ablation) | TranSVAE · UCF-HMDB average | accuracy: 60.60 -> 63.19 | Helped |
| Add contrastive triplet loss for static domain-specific factors (ablation) | TranSVAE · UCF-HMDB average | accuracy: 60.64 -> 63.19 | Helped |
| Add sequential VAE reconstruction and latent-factor generation (ablation) | TranSVAE · UCF-HMDB U→H | accuracy: 82.22 -> 87.78 | Helped |
| Use T=8 rather than T=5 frames (ablation) | TranSVAE · UCF-HMDB U→H | accuracy: 84.44 -> 87.78 | Helped |
| Use target pseudo-label supervision (ablation) | TranSVAE · UCF-HMDB H→U | accuracy: 94.40 -> 98.95 | Helped |
Closest papers by meaning
- Unsupervised Video Domain Adaptation with Masked Pre-Training and Collaborative Self-Training (CVPR 2024 · similarity 0.91)
- Return of Frustratingly Easy Unsupervised Video Domain Adaptation (ICML 2026 · similarity 0.90)
- Discovering Informative and Robust Positives for Video Domain Adaptation (ICLR 2023 · similarity 0.90)
- Multi-Modal Domain Adaptation for Fine-Grained Action Recognition (CVPR 2020 · similarity 0.89)
- Modality-Collaborative Test-Time Adaptation for Action Recognition (CVPR 2024 · similarity 0.88)
- DVANet: Disentangling View and Action Features for Multi-View Action Recognition (AAAI 2024 · similarity 0.86)
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.