Research map / 3D vision / Human pose estimation
Human Mesh Recovery From Monocular Images via a Skeleton-Disentangled Representation
Yu Sun, Yun Ye, Wu Liu, Wenpeng Gao, Yili Fu, Tao Mei
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 DSD module (bilinear skeleton disentangling) (ablation) | Direct estimation (ResNet-50) · Human3.6M | MPJPE: 110.0 -> 60.1 | Improved |
| add DSD module (bilinear skeleton disentangling) (ablation) | Direct estimation (ResNet-50) · Human3.6M | PA-MPJPE: 62.8 -> 44.3 | Improved |
| concatenate skeleton and detail features instead of bilinear transformation (ablation) | Direct estimation · Human3.6M | PA-MPJPE: 62.8 -> 63.1 | Hurt |
| use bilinear transformation instead of concatenation (ablation) | Concat · Human3.6M | MPJVE: 20.5 -> 12.5 | Improved |
| add SATN temporal network (ablation) | DSD · Human3.6M | PA-MPJPE: 44.3 -> 42.4 | Improved |
| add SATN temporal network (ablation) | DSD · 3DPW test | PA-MPJPE: 75.0 -> 69.5 | Improved |
| add SATN temporal network (ablation) | DSD · Human3.6M | MPJVE: 12.5 -> 8.7 | Improved |
| add TCN temporal modeling alone (ablation) | DSD · Human3.6M | PA-MPJPE: 44.3 -> 52.0 | Hurt |
| add self-attention before the TCN (ablation) | DSD+TCN · Human3.6M | PA-MPJPE: 52.0 -> 43.4 | Improved |
| add adversarial temporal-order training (ablation) | DSD+TCN+Self-attention · Human3.6M | PA-MPJPE: 43.4 -> 42.4 | Improved |
| add DSD module under TCN temporal modeling (ablation) | Backbone+TCN · Human3.6M | PA-MPJPE: 58.5 -> 52.0 | Improved |
| replace TCN with SATN without DSD (ablation) | Backbone+TCN · Human3.6M | PA-MPJPE: 58.5 -> 58.6 | No clear effect |
Closest papers by meaning
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.89)
- Self-Supervised 3D Human Mesh Recovery from a Single Image with Uncertainty-Aware Learning (AAAI 2024 · similarity 0.89)
- Reconstructing Humans with a Biomechanically Accurate Skeleton (CVPR 2025 · similarity 0.88)
- MultiGO: Towards Multi-level Geometry Learning for Monocular 3D Textured Human Reconstruction (CVPR 2025 · similarity 0.88)
- Disentangled Diffusion-Based 3D Human Pose Estimation with Hierarchical Spatial and Temporal Denoiser (AAAI 2024 · similarity 0.88)
- D^3-Human: Dynamic Disentangled Digital Human from Monocular Video (CVPR 2025 · similarity 0.88)
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.
