Research map / 3D vision / Face and shape
Monocular Real-Time Full Body Capture With Inter-Part Correlations
Check 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 |
|---|---|---|---|
| stream body-features into hand detection and combine them with local supplementary features (ablation) | DetNet hand branch · MTC-Hand | MPJPE (mm): 18.4 -> 14.4; 28% lower error than DetNet-S | Helped |
| add body-features to a hand branch using local supplementary features (ablation) | DetNet hand branch · MTC-Hand-Mask | MPJPE (mm): 31.7 -> 30.6; gap ... -5% | No clear effect |
| add body-features to hand detection (ablation) | DetNet hand branch · FreiHand | MPJPE (mm): 23.1 -> 24.2; gap ... 4% | Mixed |
| remove high-resolution supplementary hand features (ablation) | DetNet hand branch · MTC-Hand-Mask | MPJPE (mm): 30.6 -> 37.5 | Hurt |
| remove high-resolution supplementary hand features (ablation) | DetNet hand branch · FreiHand | MPJPE (mm): 24.2 -> 26.8 | Hurt |
| remove the attention mechanism by treating hand-only examples as if body is present (ablation) | DetNet training · MPII3D | MPJPE (mm): 90.1 -> 99.9 | Hurt |
| train only on MTC, the dataset with simultaneous body and hand annotations (ablation) | DetNet training · HM36M | MPJPE (mm): 57.5 -> 272.2 | Hurt |
| exclude 2D-labeled, in-the-wild training data (ablation) | DetNet training · HM36M | MPJPE (mm): 57.5 -> 61.7 | Hurt |
| add BodyIKNet to keypoint predictions | DetNet on HUMBI · HUMBI | MPJPE: 43.5 -> 43.3 | Helped |
| use ground-truth rather than estimated body shape parameters (ablation) | DetNet + IKNet on HUMBI · HUMBI | MPJPE: 43.3 -> 39.9 | Helped |
| use tighter face crops (ablation) | FaceNet · MTC-Face | landmark error (pixels): 3.43 -> 3.37 | Helped |
| use tighter face crops (ablation) | FaceNet · MTC-Face | photometric error per channel: 0.0447 -> 0.0444 | Helped |
Closest papers by meaning
- Monocular Total Capture: Posing Face, Body, and Hands in the Wild (CVPR 2019 · similarity 0.91)
- Expressive Body Capture: 3D Hands, Face, and Body From a Single Image (CVPR 2019 · similarity 0.89)
- Monocular Real-Time Hand Shape and Motion Capture Using Multi-Modal Data (CVPR 2020 · similarity 0.89)
- Egocentric Whole-Body Motion Capture with FisheyeViT and Diffusion-Based Motion Refinement (CVPR 2024 · similarity 0.86)
- AiOS: All-in-One-Stage Expressive Human Pose and Shape Estimation (CVPR 2024 · similarity 0.86)
- MUC: Mixture of Uncalibrated Cameras for Robust 3D Human Body Reconstruction (AAAI 2025 · 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.