Research map / Label-efficient and robust learning / Test-time adaptation
Test-Time Personalization with a Transformer for Human Pose Estimation
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 test-time personalization to Transformer keypoint training | Transformer (keypoint) · Human 3.6M | mPCK: 87.90 -> 91.70 | Helped |
| Add test-time personalization to Transformer keypoint training | Transformer (keypoint) · Penn Action | mPCK: 86.16 -> 88.98 | Helped |
| Add test-time personalization to Transformer keypoint training | Transformer (keypoint) · BBC Pose | mAcc: 90.19 -> 92.51 | Helped |
| Replace rotation prediction with self-supervised keypoint estimation (ablation) | Feature-shared joint training · Human 3.6M | mPCK: 87.37 -> 87.41 | Helped |
| Replace rotation prediction with self-supervised keypoint estimation (ablation) | Feature-shared joint training · Penn Action | mPCK: 84.90 -> 85.78 | Helped |
| Replace rotation prediction with self-supervised keypoint estimation (ablation) | Feature-shared joint training · BBC Pose | mAcc: 89.07 -> 89.65 | Helped |
| Add Transformer affinity mapping between self-supervised and supervised keypoints (ablation) | Feature-shared keypoint model · Human 3.6M | mPCK: 87.41 -> 87.90 | Helped |
| Add Transformer affinity mapping between self-supervised and supervised keypoints (ablation) | Feature-shared keypoint model · Penn Action | mPCK: 85.78 -> 86.16 | Helped |
| Add Transformer affinity mapping between self-supervised and supervised keypoints (ablation) | Feature-shared keypoint model · BBC Pose | mAcc: 89.65 -> 90.19 | Helped |
| Use offline instead of online personalization (ablation) | Transformer (keypoint) · BBC Pose | mAcc: 92.51 -> 92.21 | Hurt |
| Use offline instead of online personalization (ablation) | Transformer (keypoint) · Human 3.6M | mPCK: 91.70 -> 92.05 | Helped |
| Use rotation prediction as the self-supervised objective (ablation) | Feature-shared rotation model · Human 3.6M | mPCK: 85.42 -> 84.90 | Hurt |
Closest papers by meaning
- SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estimation (CVPR 2024 · similarity 0.87)
- Fast Adaptation for Human Pose Estimation via Meta-Optimization (CVPR 2024 · similarity 0.87)
- SDPose: Tokenized Pose Estimation via Circulation-Guide Self-Distillation (CVPR 2024 · similarity 0.86)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.85)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (CVPR 2025 · similarity 0.85)
- Personalized Representation from Personalized Generation (ICLR 2025 · 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.