Research map / Video / Video segmentation and tracking
Learning 3D Human Dynamics From Video
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 |
|---|---|---|---|
| Use temporal context rather than a single image | HMMR · 3DPW | MPJPE: 130.0 -> 127.1 | Helped |
| Use temporal context rather than a single image | HMMR · 3DPW | Accel Error: 37.4 -> 16.4 | Helped |
| Add auxiliary past and future pose-change prediction (ablation) | HMMR temporal encoder · 3DPW | MPJPE: 139.2 -> 127.1 | Helped |
| Add auxiliary past and future pose-change prediction (ablation) | HMMR temporal encoder · Penn Action | PCK: 71.2 -> 77.9 | Helped |
| Add auxiliary past and future pose-change prediction (ablation) | HMMR temporal encoder · 3DPW | Accel Error: 15.2 -> 16.4 | Mixed |
| Add constant-shape loss (ablation) | HMMR · 3DPW | PA-MPJPE: 78.2 -> 80.1 | Helped |
| Add constant-shape loss (ablation) | HMMR · Human3.6M | MPJPE: 83.5 -> 87.0 | Mixed |
| Add VLOG pseudo-ground-truth 2D video | HMMR · 3DPW | MPJPE: 127.1 -> 126.7 | Helped |
| Add InstaVariety pseudo-ground-truth 2D video | HMMR · 3DPW | MPJPE: 127.1 -> 116.5 | Helped |
| Add InstaVariety pseudo-ground-truth 2D video | HMMR · NBA | PCK: 68.4 -> 68.1 | Mixed |
| Add InstaVariety pseudo-ground-truth 2D video | HMMR · Human3.6M | PA-MPJPE: 58.1 -> 56.9 | Helped |
| Use temporal context instead of local single-view input | HMMR · Penn Action | PCK: 73.2 -> 77.9 | Helped |
Closest papers by meaning
- VideoMDM: Towards 3D Human Motion Generation From 2D Supervision (NeurIPS 2026 · similarity 0.91)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.90)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.89)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.88)
- Lifting Motion to the 3D World via 2D Diffusion (CVPR 2025 · similarity 0.88)
- S3O: A Dual-Phase Approach for Reconstructing Dynamic Shape and Skeleton of Articulated Objects from Single Monocular Video (ICML 2024 · 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.