Research map / Detection and segmentation / CNN features
Learning Quality-Aware Representation for Multi-Person Pose Regression
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 CIR to the Gaussian-label baseline (ablation) | Single-stage pose regression with HRNet-W32 · COCO mini-val | AP: 63.5 -> 64.5 | Helped |
| Add KQE to CIR (ablation) | Single-stage pose regression with HRNet-W32 · COCO mini-val | AP: 64.5 -> 68.4 | Helped |
| Add PQE to CIR + KQE (ablation) | Single-stage pose regression with HRNet-W32 · COCO mini-val | AP: 68.4 -> 69.8 | Helped |
| Add CIR to KQE + PQE (ablation) | Single-stage pose regression with HRNet-W32 · COCO mini-val | AP: 68.3 -> 69.8 | Helped |
| Replace CIR with discrete {0,1} instance representation (ablation) | Pose regression with QEB disabled · COCO mini-val | AP: 69.8 -> 67.7 | Hurt |
| Add a quality estimation branch to discrete {0,1} representation (ablation) | Single-stage pose regression · COCO mini-val | AP: 67.7 -> 68.6 | Helped |
| Add a quality estimation branch to Gaussian representation (ablation) | Single-stage pose regression · COCO mini-val | AP: 68.3 -> 69.0 | Helped |
| Set KQE semantic-point count to 9 instead of 0 (ablation) | KQE with the other proposed components fixed · COCO mini-val | AP: 68.5 -> 69.8 | Helped |
| Vary the KQE semantic-point count from 9 to 12 or 15 (ablation) | KQE with the other proposed components fixed · COCO mini-val | AP: 69.8 -> 69.6; 69.8 -> 69.1 | Hurt |
| Aggregate pose-query features at keypoint queries rather than automatic points, semantic points, or keypoint positions (ablation) | Pose Query Encoding · COCO mini-val | AP: 68.8; 68.9; 69.0 -> 69.8 | Helped |
| Use the proposed method instead of DEKR-W48 | COCO pose estimation with HRNet-W48, single-scale testing · COCO test-dev | AP: 70.0 -> 71.0 | Helped |
| Use the proposed method instead of DEKR-W48 with multi-scale testing | COCO pose estimation with HRNet-W48 · COCO test-dev | AP: 71.0 -> 71.7 | Helped |
Closest papers by meaning
- QueryPose: Sparse Multi-Person Pose Regression via Spatial-Aware Part-Level Query (NeurIPS 2022 · similarity 0.90)
- Simple Pose: Rethinking and Improving a Bottom-up Approach for Multi-Person Pose Estimation (AAAI 2020 · similarity 0.87)
- On the Calibration of Human Pose Estimation (ICML 2024 · similarity 0.85)
- PoseBH: Prototypical Multi-Dataset Training Beyond Human Pose Estimation (CVPR 2025 · similarity 0.85)
- TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose Representation (CVPR 2024 · similarity 0.85)
- DiffusionRegPose: Enhancing Multi-Person Pose Estimation using a Diffusion-Based End-to-End Regression Approach (CVPR 2024 · 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.