Research map / Video / Action recognition
Towards Omni-Supervised Face Alignment for Large Scale Unlabeled Videos
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 |
|---|---|---|---|
| Add spatial-temporal relational reasoning to MDM | MDM · 300VW Category 1 | RMSE: 9.88 -> 5.03 | Helped |
| Add spatial-temporal relational reasoning to MDM | MDM · 300VW Category 2 | RMSE: 8.14 -> 4.74 | Helped |
| Add spatial-temporal relational reasoning to MDM | MDM · 300VW Category 3 | RMSE: 9.81 -> 6.63 | Helped |
| Add spatial-temporal relational reasoning to HGN | HGN · 300VW Category 1 | RMSE: 6.12 -> 5.31 | Helped |
| Add spatial-temporal relational reasoning to HGN | HGN · 300VW Category 2 | RMSE: 5.33 -> 4.77 | Helped |
| Add spatial-temporal relational reasoning to HGN | HGN · 300VW Category 3 | RMSE: 7.04 -> 8.06 | Hurt |
| Add YouTube-Face videos to training | STRRN · 300VW+YouTube-Face | AUC@0.05: 26.67 -> 36.33 | Helped |
| Add YouTube-Face videos to training | STRRN · 300VW+YouTube-Face | AUC@0.08: 78.07 -> 80.43 | Helped |
| Add YouTube-Celebrities videos to training | STRRN · 300VW+YouTube-Celebrities | AUC@0.05: 26.67 -> 28.31 | Helped |
| Add YouTube-Celebrities videos to training | STRRN · 300VW+YouTube-Celebrities | AUC@0.08: 78.07 -> 77.24 | Hurt |
| Add cycle-consistency loss (ablation) | STRRN · 300VW Category 3 | RMSE: 8.78 -> 6.63 | Helped |
| Add detector loss (ablation) | STRRN · 300VW Category 2 | RMSE: 7.08 -> 4.74 | Helped |
Closest papers by meaning
- Dynamic Facial Analysis: From Bayesian Filtering to Recurrent Neural Network (CVPR 2017 · similarity 0.85)
- Temporal Alignment Networks for Long-Term Video (CVPR 2022 · similarity 0.84)
- Self-Supervised Facial Representation Learning with Facial Region Awareness (CVPR 2024 · similarity 0.83)
- Omni-RGPT: Unifying Image and Video Region-level Understanding via Token Marks (CVPR 2025 · similarity 0.83)
- PrefAce: Face-Centric Pretraining with Self-Structure Aware Distillation (AAAI 2024 · similarity 0.83)
- ODTrack: Online Dense Temporal Token Learning for Visual Tracking (AAAI 2024 · similarity 0.82)
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.