Research map / 3D vision / Human pose estimation
Continuous-Time Human Motion Field from Event Cameras
Ziyun Wang, Ruijun Zhang, Zi-Yan Liu, Yufu Wang, Kostas Daniilidis
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 fine-tuning of global motion predictor (ablation) | Ours · High-speed event human dataset (15 FPS eval) | MPJPE: 67.01 -> 49.74 | Improved |
| add fine-tuning of global motion predictor (ablation) | Ours · High-speed event human dataset (15 FPS eval) | PA-MPJPE: 39.16 -> 30.05 | Improved |
| add fine-tuning of global motion predictor (ablation) | Ours · High-speed event human dataset (15 FPS eval) | PEL-MPJPE: 52.23 -> 41.06 | Improved |
| add fine-tuning of global motion predictor (ablation) | Ours · High-speed event human dataset (15 FPS eval) | PCKh@0.5: 0.86 -> 0.92 | Improved |
| add event contrast maximization loss (ablation) | Ours · High-speed event human dataset (15 FPS eval) | MPJPE: 50.77 -> 49.74 | Improved |
| add event contrast maximization loss (ablation) | Ours · High-speed event human dataset (15 FPS eval) | PA-MPJPE: 30.22 -> 30.05 | Improved |
| add event contrast maximization loss (ablation) | Ours · High-speed event human dataset (15 FPS eval) | PCKh@0.5: 0.91 -> 0.92 | Improved |
| replace EventHPE with continuous-time motion field | Ours · MMHPSD | PA-MPJPE: 43.90 -> 39.16 | Improved |
| replace EventHPE with continuous-time motion field | Ours · High-speed event human dataset (15 FPS eval) | MPJPE: 15.54 mm | Improved |
| replace EventHPE with continuous-time motion field | Ours · High-speed event human dataset (120 FPS eval) | MPJPE: 66.76 -> 49.76 | Improved |
| use continuous-time decoding instead of per-frame prediction (computation time) | Ours · long prediction sequences | computation time: 69% | Improved |
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.
- EventHPE — EventHPE: Event-Based 3D Human Pose and Shape Estimation (ICCV 2021) · better on 6 of 6 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
- Natural Human Motion Recovery by Aligning High-Order Temporal Dynamics from Monocular Videos (CVPR 2026 · similarity 0.89)
- Humans as a Calibration Pattern: Dynamic 3D Scene Reconstruction from Unsynchronized and Uncalibrated Videos (ICCV 2025 · similarity 0.89)
- Contact and Human Dynamics from Monocular Video (ECCV 2020 · similarity 0.88)
- EventCap: Monocular 3D Capture of High-Speed Human Motions Using an Event Camera (CVPR 2020 · similarity 0.88)
- HumanMM: Global Human Motion Recovery from Multi-shot Videos (CVPR 2025 · similarity 0.87)
- Optimal-state Dynamics Estimation for Physics-based Human Motion Capture from Videos (NeurIPS 2024 · 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.
