Research map / 3D vision / Human pose estimation
Divide and Fuse: Body Part Mesh Recovery from Partially Visible Human Images
Tianyu Luan, Zhongpai Gao, Luyuan Xie, Abhishek Sharma, Hao Ding, Benjamin Planche, Meng Zheng, Ange Lou, Terrence Chen, Junsong Yuan, Ziyan Wu
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 6D rotation loss on HPPM part rotations (ablation) | D&F · PV-Human3.6M (finetuned) | MPVE (mm): 95.6 -> 63.3 | Improved |
| add overlapping loss between adjacent parts (ablation) | D&F · PV-Human3.6M (finetuned) | MPVE (mm): 74.5 -> 63.3 | Improved |
| add translation loss on part global translation (ablation) | D&F · PV-Human3.6M (finetuned) | MPVE (mm): 75.1 -> 63.3 | Improved |
| add part 3D per-vertex loss (ablation) | D&F · PV-3DPW (finetuned) | MPVE (mm): 120.2 -> 109.9 | Improved |
| add depth consistency loss (ablation) | D&F · PV-Human3.6M (finetuned) | MPVE (mm): 65.4 -> 63.3 | Improved |
| add part 2D projection loss (ablation) | D&F · PV-Human3.6M (finetuned) | MPVE (mm): 64.2 -> 63.3 | Improved |
| use adjustable number of shape parameters per part instead of fixed 24 per part (ablation) | D&F · PV-Human3.6M (finetuned) | MPVE (mm): 67.5 -> 63.3 | Improved |
| add partially visible augmentation finetuning | D&F · PV-Human3.6M | MPVE (mm): 155.7 -> 63.3 | Improved |
| replace top-down SMPL-based methods with bottom-up part-based D&F | partially visible human mesh recovery · PV-3DPW (directly tested) | MPVE (mm): 268.2 -> 208.6 | Improved |
| increase number of visible parts from 1-4 to 5-10 | D&F · PV-3DPW | MPVE (mm): 109.9 -> 117.4 | Mixed |
| apply D&F to full-body visible inputs without modification | D&F · 3DPW (full-body visible) | MPJPE (mm): 100.5 vs VIBE 93.5 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- VIBE — VIBE: Video Inference for Human Body Pose and Shape Estimation (CVPR 2020) · better on 1 of 2 shared results
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
- Recovering 3D Human Mesh From Monocular Images: A Survey (TPAMI 2023 · similarity 0.90)
- MeshPose: Unifying DensePose and 3D Body Mesh Reconstruction (CVPR 2024 · similarity 0.88)
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.88)
- BLADE: Single-view Body Mesh Estimation through Accurate Depth Estimation (CVPR 2025 · similarity 0.87)
- Self-Supervised 3D Human Mesh Recovery from a Single Image with Uncertainty-Aware Learning (AAAI 2024 · similarity 0.87)
- PSHuman: Photorealistic Single-image 3D Human Reconstruction using Cross-Scale Multiview Diffusion and Explicit Remeshing (CVPR 2025 · similarity 0.87)
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.
