Research map / 3D vision / Human pose estimation
We Are More Than Our Joints: Predicting How 3D Bodies Move
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 |
|---|---|---|---|
| add latent DCT space to the CVAE | VAE+DCT+5 versus DLow · Human3.6M | ADE: 0.425 -> 0.412 | Helped |
| sample from the lowest 5 frequency bands with DLow rather than standard normal across all bands (ablation) | VAE+DCT · Human3.6M | diversity: 3.462 -> 12.579 | Helped |
| remove latent DCT space (ablation) | MOJO-proj · BMLHandball | ADE: 1.528 -> 1.608 | Hurt |
| remove latent DCT space (ablation) | MOJO · ACCAD | foot skate: 0.278 -> 0.341 | Hurt |
| add recursive projection by fitting SMPL-X at each prediction step (ablation) | joint-based representation · ACCAD | BDF: 0.185 -> 0 | Helped |
| use 67 SSM2 markers instead of 41 CMU markers (ablation) | marker-based representation · ACCAD | ADE: 1.214 -> 1.124 | Helped |
| use 67 SSM2 markers instead of 41 CMU markers (ablation) | marker-based representation · BMLHandball | ADE: 0.949 -> 0.924 | Helped |
| add joints to the marker-based representation (ablation) | CMU 41 marker representation · BMLHandball | ADE: 0.949 -> 0.943 | No clear effect |
| increase input frames and shorten the predicted sequence (ablation) | MOJO · ACCAD | diversity: 32.666 -> 3.728 | Hurt |
| increase input frames and shorten the predicted sequence (ablation) | MOJO · ACCAD | ADE: 1.303 -> 0.642 | Helped |
| remove the recursive projection scheme (ablation) | MOJO · ACCAD | upper torso deformation (mm): 32.7 -> 80.7 | Hurt |
| increase latent dimension from 16d to 128d (ablation) | MOJO-proj · BMLHandball | diversity: 23.660 -> 25.265 | Helped |
Closest papers by meaning
- Forecasting of 3D Whole-body Human Poses with Grasping Objects (CVPR 2024 · similarity 0.90)
- Expressive Forecasting of 3D Whole-Body Human Motions (AAAI 2024 · similarity 0.89)
- Nonisotropic Gaussian Diffusion for Realistic 3D Human Motion Prediction (CVPR 2025 · similarity 0.89)
- SimpliHuMoN: Simplifying Human Motion Prediction (NeurIPS 2026 · similarity 0.88)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.88)
- Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation (NeurIPS 2024 · 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.