Research map / Detection and segmentation / CNN features
QueryPose: Sparse Multi-Person Pose Regression via Spatial-Aware Part-Level Query
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 SPEGM to regression baseline (ablation) | QueryPose R50 · COCO mini-val | AP: 59.2 -> 63.8 | Helped |
| add SIM on top of SPEGM (ablation) | QueryPose R50 · COCO mini-val | AP: 63.8 -> 64.9 | Helped |
| use part-level queries instead of instance-level query (MHSA interaction) (ablation) | QueryPose R50 · COCO mini-val | AP: 60.7 -> 62.6 | Helped |
| SPEGM part embeddings instead of instance-level spatial attention (ablation) | QueryPose R50 · COCO mini-val | AP: 60.2 -> 63.8 | Helped |
| selective iteration instead of summation (ablation) | QueryPose R50 · COCO mini-val | AP: 64.2 -> 64.9 | Helped |
| selective iteration instead of concatenation (ablation) | QueryPose R50 · COCO mini-val | AP: 64.1 -> 64.9 | Helped |
| increase part query dimension from 128 to 256 (ablation) | QueryPose R50 · COCO mini-val | AP: 64.9 -> 64.7 | No clear effect |
| part division by rigid parts (7) instead of per-keypoint (17) (ablation) | QueryPose R50 · COCO mini-val | AP: 64.7 -> 64.9 | Helped |
| iterate instance query across box and keypoint decoders serially (ablation) | QueryPose R50 · COCO mini-val | AP_box: 55.3 -> 56.9 | Helped |
| sparse part-level queries instead of dense end-to-end methods | QueryPose HRNet-W48 · COCO mini-val | AP: 70.8 -> 73.6 | Helped |
| add online auxiliary keypoint heatmap learning | QueryPose · COCO | AP | No clear effect |
Closest papers by meaning
- Learning Quality-Aware Representation for Multi-Person Pose Regression (AAAI 2022 · similarity 0.90)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (CVPR 2025 · similarity 0.89)
- MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation (CVPR 2025 · similarity 0.87)
- PoseBH: Prototypical Multi-Dataset Training Beyond Human Pose Estimation (CVPR 2025 · similarity 0.86)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.86)
- SHaRPose: Sparse High-Resolution Representation for Human Pose Estimation (AAAI 2024 · 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.