Research map / Label-efficient and robust learning / Semi-supervised learning
Semi-Supervised 2D Human Pose Estimation Driven by Position Inconsistency Pseudo Label Correction Module
Linzhi Huang, Yulong Li, Hongbo Tian, Yue Yang, Xiangang Li, Weihong Deng, Jieping Ye
AI-extracted. The measured effects below were read by AI from the paper's full text (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 PCM pseudo-label correction (ablation) | ResNet18 pose estimator · COCO 10K labeled | AP: 51.1 -> 59.6 | Improved |
| Add SSCO to PCM (ablation) | ResNet18 pose estimator · COCO 10K labeled | AP: 59.6 -> 60.7 | Improved |
| Add SSCO to DUAL (ablation) | ResNet18 pose estimator · COCO 10K labeled | AP: 58.7 -> 60.1 | Improved |
| Use SSCO instead of Cutout, Mixup, CutMix, Rand Augment, or JC with PCM (ablation) | ResNet18 pose estimator · COCO 10K labeled | AP: 60.0, 58.2, 60.5, 59.9, 60.0 -> 60.7 | Mixed |
| Use 2 local limb images in SSCO rather than 0, 1, 3, or 4 (ablation) | PCM + SSCO · COCO 10K labeled | AP: 59.6, 60.2, 60.4, 60.5 -> 60.7 | Improved |
| Use SSPCM rather than DUAL | ResNet18 pose estimator · COCO 1K labeled | AP: 44.6 -> 46.9 | Improved |
| Use SSPCM rather than DUAL | ResNet18 pose estimator · COCO 5K labeled | AP: 55.6 -> 57.5 | Improved |
| Use SSPCM rather than DUAL | ResNet18 pose estimator · COCO 10K labeled | AP: 59.6 -> 60.7 | Improved |
| Use a ResNet50 teacher with a ResNet18 student rather than DUAL with the same teacher/student pair | ResNet18 student · COCO 10K labeled | AP: 60.4 -> 61.9 | Improved |
| Use SSPCM rather than Dual | Pose estimator · WEPDTOF-Pose TEST, CEPDOF unlabeled | AP: 55.1 -> 55.6 | Improved |
| Use SSPCM rather than Dual | Pose estimator · WEPDTOF-Pose TEST, BKFisheye UNLABEL | AP: 57.2 -> 59.1 | Improved |
| Use SSPCM rather than Dual | Pose estimator · BKFisheye TEST, BKFisheye UNLABEL | AP: 68.4 -> 68.7 | Improved |
Closest papers by meaning
- An Empirical Study of the Collapsing Problem in Semi-Supervised 2D Human Pose Estimation (ICCV 2021 · similarity 0.86)
- DynPose: Largely Improving the Efficiency of Human Pose Estimation by a Simple Dynamic Framework (CVPR 2025 · similarity 0.84)
- Pose Estimation Errors, the Ultimate Diagnosis (ECCV 2016 · similarity 0.84)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (CVPR 2025 · similarity 0.82)
- Sequential Joint Dependency Aware Human Pose Estimation with State Space Model (AAAI 2025 · similarity 0.82)
- SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estimation (CVPR 2024 · similarity 0.81)
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.
