Research map / 3D vision / Human pose estimation
3D Hand Pose Estimation in Everyday Egocentric Images
Aditya Prakash, Ruisen Tu, Matthew Chang, Saurabh Gupta
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 |
|---|---|---|---|
| use hand crops as input instead of full image | ArcticNet-SF · H2O (zero-shot) | MPJPE (mm): 83.84 -> 38.06 | Improved |
| add KPE (camera-FOV location encoding) to crop-based model (ablation) | HandNet · H2O (zero-shot) | MRRPE (mm): 141.06 -> 49.49 | Improved |
| add KPE (ablation) | HandNet · EpicHands (zero-shot) | L2 error (pixels): 31.62 -> 11.05 | Improved |
| add KPE to HaMeR ViT | HaMeR (ViT) · EpicHands (zero-shot) | L2 error (pixels): 25.48 -> 9.77 | Improved |
| add KPE to HandOccNet | HandOccNet · H2O (zero-shot) | MPJPE (mm): 60.58 -> 47.57 | Improved |
| add hand mask supervision from in-the-wild data (ablation) | WildHands · H2O (zero-shot) | MPJPE (mm): 39.52 -> 34.29 | Improved |
| add grasp supervision from in-the-wild data (ablation) | WildHands · H2O (zero-shot) | MPJPE (mm): 39.52 -> 38.34 | Improved |
| add Ego4D grasp supervision alone (ablation) | WildHands · H2O (zero-shot) | MRRPE (mm): 77.07 -> 111.47 | Hurt |
| use sparse KPE instead of dense KPE (ablation) | WildHands · H2O (zero-shot) | MPJPE (mm): 36.86 -> 31.08 | Improved |
| add KPE to the input instead of latent features (ablation) | WildHands · H2O (zero-shot) | MPJPE (mm): 38.06 -> 45.51 | Hurt |
| use KPE instead of CamConv, perspective correction or PCL (ablation) | HandNet · EpicHands (zero-shot) | L2 error (pixels): 17.35 (CamConv) -> 11.05 | Improved |
| add ASSEMBLY and Ego4D data to ARCTIC training (ablation) | WildHands · EpicHands (zero-shot) | L2 error (pixels): 17.07 -> 7.20 | 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.
- HandOccNet — HandOccNet: Occlusion-Robust 3D Hand Mesh Estimation Network (CVPR 2022) · better on 1 of 1 shared result
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
- Benchmarks and Challenges in Pose Estimation for Egocentric Hand Interactions with Objects (ECCV 2024 · similarity 0.90)
- AssemblyHands: Towards Egocentric Activity Understanding via 3D Hand Pose Estimation (CVPR 2023 · similarity 0.90)
- HOT3D: Hand and Object Tracking in 3D from Egocentric Multi-View Videos (CVPR 2025 · similarity 0.89)
- Single-to-Dual-View Adaptation for Egocentric 3D Hand Pose Estimation (CVPR 2024 · similarity 0.89)
- EgoPressure: A Dataset for Hand Pressure and Pose Estimation in Egocentric Vision (CVPR 2025 · similarity 0.87)
- Analyzing the Synthetic-to-Real Domain Gap in 3D Hand Pose Estimation (CVPR 2025 · 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.
