Research map / 3D vision / Human pose estimation
Heatmap Distribution Matching 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 |
|---|---|---|---|
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | Simple Baseline (ResNet-152) · COCO val | AP: 75.0 -> 76.7 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRNet-W32 · COCO val | AP: 76.7 -> 78.2 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRNet-W48 · COCO val | AP: 77.1 -> 78.8 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRFormer-Base · COCO val | AP: 78.0 -> 78.9 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | Simple Baseline (ResNet-152) · COCO test-dev | AP: 73.8 -> 75.3 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRNet-W48 · COCO test-dev | AP: 75.5 -> 77.2 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRFormer-Base · COCO test-dev | AP: 76.2 -> 77.2 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | Simple Baseline (ResNet-152) · MPII val | PCKh@0.5 Mean: 89.6 -> 90.3 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRNet-W32 · MPII val | PCKh@0.5 Mean: 90.4 -> 90.9 | Helped |
| Replace Gaussian-smoothed heatmap pixel-wise loss with distribution matching to the dot annotation | HRNet-W48 · MPII val | PCKh@0.5 Mean: 90.5 -> 90.9 | Helped |
| Use four sub-pixel demanders instead of a single demander at the containing pixel center (ablation) | Distribution matching with HRNet-W48 · COCO val | AP: 77.9 -> 78.8 | Helped |
| Use 500 instead of 1000 Sinkhorn iterations (ablation) | Distribution matching with HRNet-W48 · COCO val | AP: 78.8 -> 78.3 | Hurt |
Closest papers by meaning
- Continuous Heatmap Regression for Pose Estimation via Implicit Neural Representation (NeurIPS 2024 · similarity 0.88)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · similarity 0.87)
- MotionMap: Representing Multimodality in Human Pose Forecasting (CVPR 2025 · similarity 0.87)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (CVPR 2025 · similarity 0.86)
- Video-Based Human Pose Regression via Decoupled Space-Time Aggregation (CVPR 2024 · similarity 0.86)
- Attention-Propagation Network for Egocentric Heatmap to 3D Pose Lifting (CVPR 2024 · similarity 0.84)
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.