Research map / 3D vision / Human pose estimation
Uni-Hand: Universal Hand Motion Forecasting in Egocentric Views
Junyi Ma, Wentao Bao, Jingyi Xu, G. X. Sun, Y. Zheng, Erhang Zhang, Xieyuanli Chen, Hesheng Wang
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 target indicators X^tar (ablation) | Uni-Hand · H2O-PT multi-joint (3D) | ADE: 0.038 -> 0.036 | Improved |
| add target indicators X^tar (ablation) | Uni-Hand · HOT3D-Clips multi-joint (3D) | ADE: 0.121 -> 0.118 | Improved |
| add target indicators X^tar (ablation) | Uni-Hand · CABH-E multi-joint (3D) | ADE: 0.047 -> 0.038 | Improved |
| add target indicators X^tar (ablation) | Uni-Hand · CABH-E multi-joint (3D) | FDE: 0.056 -> 0.046 | Improved |
| add predicted HM features to RU-LSTM RGB branch | RU-LSTM · Epic-Kitchens-55 action recognition (100%) | Top-5 accuracy: 41.78 -> 42.74 | Improved |
| add predicted HM features to RU-LSTM flow branch | RU-LSTM · Epic-Kitchens-55 action recognition (100%) | Top-5 accuracy: 30.96 -> 31.74 | Improved |
| use multi-joint HM features instead of center-only HM features (ablation) | RU-LSTM RGB · Epic-Kitchens-55 action recognition (100%) | Top-5 accuracy: 42.74 -> 42.82 | Improved |
| use multi-joint HM features instead of center-only HM features (ablation) | RU-LSTM flow · Epic-Kitchens-55 action recognition (100%) | Top-5 accuracy: 31.74 -> 31.86 | Improved |
| add predicted HM features to RU-LSTM RGB branch | RU-LSTM · Epic-Kitchens-55 early action recognition | Top-5 accuracy: 36.40 -> 36.22 at 12.5%; 41.66 -> 42.32 at 87.5% | Mixed |
| add predicted HM features to RU-LSTM flow branch | RU-LSTM · Epic-Kitchens-55 early action recognition | Top-5 accuracy: 27.63 -> 26.90 at 37.5%; 30.71 -> 31.74 at 87.5% | Mixed |
| replace S-Mamba with Uni-Hand | HAT robot policy · HAT pick-and-place task | success rate: 20% -> 90% | Improved |
| replace S-Mamba with Uni-Hand | HAT robot policy · HAT harder pick-and-place task | success rate: 0% -> 80% | Improved |
Closest papers by meaning
- MEgoHand: Multimodal Egocentric Hand-Object Interaction Motion Generation (NeurIPS 2025 · similarity 0.89)
- HandEdit: A Unified Benchmark for Egocentric Human-to-Robot Dexterous Hand Image Editing (NeurIPS 2026 · similarity 0.89)
- Forecasting of 3D Whole-body Human Poses with Grasping Objects (CVPR 2024 · similarity 0.88)
- Hand-Centric Motion Refinement for 3D Hand-Object Interaction via Hierarchical Spatial-Temporal Modeling (AAAI 2024 · similarity 0.88)
- StableHand: Quality-Aware Flow Matching for World-Space Dual-Hand Motion Estimation from Egocentric Video (NeurIPS 2026 · similarity 0.88)
- Expressive Forecasting of 3D Whole-Body Human Motions (AAAI 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.
