Research map / 3D vision / Human pose estimation
Measuring Generalisation to Unseen Viewpoints, Articulations, Shapes and Objects for 3D Hand Pose Estimation under Hand-Object Interaction
Anil Armagan, Guillermo Garcia-Hernando, Seungryul Baek, Shreyas Hampali, Mahdi Rad, Zhaohui Zhang, Shipeng Xie, MingXiu Chen, Boshen Zhang, Fu Xiong, Yang Xiao, Zhiguo Cao and 23 more
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 segmentation-network hand center instead of threshold center (ablation) | AWR (ResNet-50) · HANDS'19 Task 1 extrapolation | MJE (mm): 20.70 -> 14.89 | Improved |
| crop the hand from the segmented mask instead of original depth (ablation) | AWR (ResNet-50) · HANDS'19 Task 1 extrapolation | MJE (mm): 14.89 -> 14.75 | Improved |
| use ResNet-101 instead of ResNet-50 backbone (ablation) | AWR · HANDS'19 Task 1 extrapolation | MJE (mm): 14.75 -> 14.44 | Improved |
| ensemble five models (ablation) | AWR · HANDS'19 Task 1 extrapolation | MJE (mm): 14.44 -> 13.67 | Improved |
| use deeper ResNet-152 instead of ResNet-101 (ablation) | NLE · HANDS'19 Task 3 extrapolation | MJE (mm): 32.56 -> 37.56 | Hurt |
| quadruple the kernels of V2V-PoseNet (wider) (ablation) | NTIS · HANDS'19 extrapolation | MJE (mm): 38.33 -> 36.36 | Improved |
| increase number of anchor poses from 1 to 50 (ablation) | NLE · HANDS'19 Task 3 shape extrapolation | MJE (mm): 21.08 -> 16.55 | Improved |
| add joint 2D pose learning to 3D-only regression (ablation) | NLE · HANDS'19 Task 3 object extrapolation | MJE (mm): 37.31 -> 28.45 | Improved |
| use pose proposal integration instead of NMS (ablation) | NLE · HANDS'19 Task 3 extrapolation | MJE (mm): 38.08 -> 34.63 | Improved |
| ensemble 12 rotated test images instead of 1 (ablation) | NLE · HANDS'19 Task 3 shape | MJE (mm): 17.35 -> 15.79 | Improved |
| ensemble the four input-type models (ablation) | Strawberryfg · HANDS'19 Task 1 extrapolation | MJE (mm): 21.19 -> 19.63 | Improved |
| add local patch refinement and volume rendering supervision (ablation) | Strawberryfg model 4 · HANDS'19 Task 1 extrapolation | MJE (mm): 22.56 -> 21.19 | Improved |
Closest papers by meaning
- InterHand2.6M: A Dataset and Baseline for 3D Interacting Hand Pose Estimation from a Single RGB Image (ECCV 2020 · similarity 0.88)
- What's in Your Hands? 3D Reconstruction of Generic Objects in Hands (CVPR 2022 · similarity 0.87)
- Robust Hand Pose Estimation During the Interaction With an Unknown Object (ICCV 2017 · similarity 0.87)
- HOISDF: Constraining 3D Hand-Object Pose Estimation with Global Signed Distance Fields (CVPR 2024 · similarity 0.86)
- UniHOPE: A Unified Approach for Hand-Only and Hand-Object Pose Estimation (CVPR 2025 · similarity 0.86)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.85)
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.
