Research map / 3D vision / Human pose estimation
AJAHR: Amputated Joint Aware 3D Human Mesh Recovery
Hyunjin Cho, Giyun Choi, Jongwon Choi
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 tokenizer switching between amputee and non-amputee codebooks (versus amputation-only tokenizer) (ablation) | AJAHR · ITW-amputee | MPJPE (mm): 131.12 -> 129.25 | Improved |
| add tokenizer switching (versus non-amputation-only tokenizer) (ablation) | AJAHR · ITW-amputee | MPJPE (mm): 134.71 -> 129.25 | Improved |
| add tokenizer switching (versus non-amputation-only tokenizer) (ablation) | AJAHR · A3D | MPJPE (mm): 76.31 -> 73.19 | Improved |
| add keypoint heatmaps to image input of BPAC-Net (ablation) | AJAHR · ITW-amputee | MPJPE (mm): 154.55 -> 129.25 | Improved |
| add image to keypoint-only input (ablation) | AJAHR · ITW-amputee | MPJPE (mm): 146.21 -> 129.25 | Improved |
| add Gaussian noise to input keypoints (100% noise ratio) (ablation) | AJAHR · ITW-amputee | MPJPE (mm): 143.74 -> 147.41 | Hurt |
| use 320 tokens instead of 640 tokens (ablation) | AJAHR-Tokenizer · ITW-amputee | MPJPE (mm): 151.80 -> 129.25 | Improved |
| use 320 tokens instead of 160 tokens (ablation) | AJAHR-Tokenizer · ITW-amputee | MPJPE (mm): 147.91 -> 129.25 | Improved |
| add BPAC-Net to weak baseline HMR2.0 (ablation) | HMR2.0 · ITW-amputee | MPJPE (mm): 154.43 -> 176.46 | Hurt |
| train on amputee data with AJAHR instead of TokenHMR | AJAHR vs TokenHMR · ITW-amputee | MPJPE (mm): 146.12 -> 129.25 | Improved |
| use AJAHR on non-amputee data versus TokenHMR | AJAHR vs TokenHMR · 3DPW | PA-MPJPE (mm): 47.17 -> 44.94 | Improved |
| use AJAHR on 3DPW versus TokenHMR (MVE) | AJAHR vs TokenHMR · 3DPW | MVE (mm): 90.23 -> 95.26 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TokenHMR — TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose Representation (CVPR 2024) · better on 7 of 7 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
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.86)
- Reconstructing Humans with a Biomechanically Accurate Skeleton (CVPR 2025 · similarity 0.86)
- Self-Supervised 3D Human Mesh Recovery from a Single Image with Uncertainty-Aware Learning (AAAI 2024 · similarity 0.85)
- ADHMR: Aligning Diffusion-based Human Mesh Recovery via Direct Preference Optimization (ICML 2025 · similarity 0.85)
- GenHMR: Generative Human Mesh Recovery (AAAI 2025 · similarity 0.85)
- DiSRT-In-Bed: Diffusion-Based Sim-to-Real Transfer Framework for In-Bed Human Mesh Recovery (CVPR 2025 · 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.
