Research map / 3D vision / Human pose estimation
Self-Constrained Inference Optimization on Structural Groups for Human Pose Estimation
Zhehan Kan, Shuoshuo Chen, Zeng Li, Zhihai He
Read the paperCheck your idea against 137,217 papers
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 self-constrained learning (SCL) (ablation) | HRNet-W48 · COCO val2017 | AP: 76.3 -> 78.3 | Improved |
| add self-constrained optimization (SCO) on top of SCL (ablation) | HRNet-W48 + SCL · COCO val2017 | AP: 78.3 -> 79.5 | Improved |
| add SCIO to HRNet-W48 | HRNet-W48 · COCO test-dev | AP: 75.5 -> 79.2 | Improved |
| add SCIO to HRNet-W32 | HRNet-W32 · COCO test-dev | AP: 74.9 -> 78.6 | Improved |
| add SCIO to SimpleBaseline R152 | SimpleBaseline · COCO test-dev | AP: 73.7 -> 77.9 | Improved |
| add SCIO to HRNet-W48 | HRNet-W48 · COCO test-dev | AR: 80.5 -> 81.6 | No clear effect |
| use SCIO instead of MIPNet | HRNet-W48 · CrowdPose test | AP: 70.0 -> 71.5 | Improved |
| use SCIO instead of Graph-PCNN with small input | HRNet-W48 · COCO val2017 (128x96) | AP: 72.8 -> 73.7 | Improved |
| add SCIO to HRNet for left wrist | HRNet · COCO val2017 | keypoint accuracy: 0.5476 -> 0.7124 | Improved |
| add SCIO to HRNet for left ankle | HRNet · COCO val2017 | keypoint accuracy: 0.3843 -> 0.5526 | Improved |
| add SCIO to HRNet for left ear | HRNet · COCO val2017 | keypoint accuracy: 0.6637 -> 0.7987 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Graph-PCNN — Graph-PCNN: Two Stage Human Pose Estimation with Graph Pose Refinement (ECCV 2020) · better on 2 of 2 shared results
- PoseFix — PoseFix: Model-Agnostic General Human Pose Refinement Network (CVPR 2019) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Self-Correctable and Adaptable Inference for Generalizable Human Pose Estimation (CVPR 2023 · similarity 0.89)
- Learning to Reconstruct 3D Human Pose and Shape via Model-Fitting in the Loop (ICCV 2019 · similarity 0.88)
- Multi-Scale Structure-Aware Network for Human Pose Estimation (ECCV 2018 · similarity 0.88)
- Adversarial PoseNet: A Structure-Aware Convolutional Network for Human Pose Estimation (ICCV 2017 · similarity 0.87)
- 3D Human Pose Estimation with Muscles (NeurIPS 2025 · similarity 0.87)
- CHAMP: Conformalized 3D Human Multi-Hypothesis Pose Estimators (ICLR 2025 · 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.
