Research map / 3D vision / Human pose estimation
EvHandPose: Event-Based 3D Hand Pose Estimation With Sparse Supervision
Jianping Jiang, Jiahe Li, Baowen Zhang, Xiaoming Deng, Boxin Shi
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 hand flow representation (ablation) | EvHandPose · Strong light, fixed poses | 3D-MPJPE: 40.30 -> 28.98 | Improved |
| hand flow instead of ECI (ablation) | EvHandPose · Strong light, fixed poses | 3D-MPJPE: 32.16 -> 28.98 | Improved |
| hand flow instead of Matrix-LSTM (ablation) | EvHandPose · Normal scenes, fixed poses | 3D-MPJPE: 58.36 -> 19.82 | Improved |
| add weak-supervision losses over fully supervised (ablation) | EvHandPose · Normal scenes, fixed poses | 3D-MPJPE: 22.76 -> 19.82 | Improved |
| add hand-edge loss (ablation) | EvHandPose · Flash scenes, fixed poses | 3D-MPJPE: 33.52 -> 32.80 | Improved |
| add CM loss (ablation) | EvHandPose · Flash scenes, fixed poses | 3D-MPJPE: 36.42 -> 32.80 | Improved |
| add hand-edge loss (ablation) | EvHandPose · Fast motion, random poses | 2D-MPJPE: 9.00 -> 8.18 | Hurt |
| add hand-edge loss (ablation) | EvHandPose · Strong light, fixed poses | 3D-MPJPE: 29.13 -> 28.98 | No clear effect |
| train on real EvRealHands instead of EventHands synthetic data (ablation) | EvHandPose · Normal scenes, fixed poses | 3D-MPJPE: 70.13 -> 19.82 | Improved |
| train on real EvRealHands instead of InterHand2.6M synthetic events (ablation) | EvHandPose · Normal scenes, fixed poses | 3D-MPJPE: 51.01 -> 19.82 | Improved |
| EvHandPose events instead of RGB MobRecon under strong light | EvHandPose · Strong light, fixed poses | 3D-MPJPE: 40.54 -> 28.98 | Improved |
| EvHandPose events instead of RGB MobRecon in normal scenes | EvHandPose · Normal scenes, fixed poses | 3D-MPJPE: 19.28 -> 19.82 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MobRecon — MobRecon: Mobile-Friendly Hand Mesh Reconstruction From Monocular Image (CVPR 2022) · better on 4 of 6 shared results
- EventHands — EventHands: Real-Time Neural 3D Hand Pose Estimation From an Event Stream (ICCV 2021) · better on 5 of 5 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
- EventHands: Real-Time Neural 3D Hand Pose Estimation From an Event Stream (ICCV 2021 · similarity 0.92)
- SeqHAND: RGB-Sequence-Based 3D Hand Pose and Shape Estimation (ECCV 2020 · similarity 0.90)
- Prior-aware Dynamic Temporal Modeling Framework for Sequential 3D Hand Pose Estimation (ICCV 2025 · similarity 0.88)
- Complementing Event Streams and RGB Frames for Hand Mesh Reconstruction (CVPR 2024 · similarity 0.88)
- ArtiBoost: Boosting Articulated 3D Hand-Object Pose Estimation via Online Exploration and Synthesis (CVPR 2022 · similarity 0.88)
- SiMHand: Mining Similar Hands for Large-Scale 3D Hand Pose Pre-training (ICLR 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.
