Research map / 3D vision / Human pose estimation
Two Heads Are Better than One: Image-Point Cloud Network for Depth-Based 3D Hand Pose Estimation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Combine depth-image visual features with point-cloud processing | IPNet · DexYCB | MPJPE (mm): 10.60 -> 8.92 | Helped |
| Add hand-pose information to point-position and 2D visual features (ablation) | IPNet initial point-cloud feature embedding · DexYCB | MPJPE (mm): 9.40 -> 9.08 | Helped |
| Add 2D visual features to point-position and hand-pose features (ablation) | IPNet initial point-cloud feature embedding · DexYCB | MPJPE (mm): 9.12 -> 9.08 | Helped |
| Add local feature aggregation (ablation) | IPNet iterative correction · DexYCB | MPJPE (mm): 9.14 -> 8.92 | Helped |
| Add GCN-based long-range interaction (ablation) | IPNet iterative correction · DexYCB | MPJPE (mm): 9.04 -> 8.92 | Helped |
| Use three correction stages instead of one (ablation) | IPNet · DexYCB | MPJPE (mm): 8.92 -> 8.71 | Helped |
| Add the 3D hand model during iterative correction (ablation) | IPNet · DexYCB | MPJPE (mm): 8.71 -> 8.36 | Helped |
| Increase PointMLP network width and number of layers (ablation) | PointMLP · DexYCB | MPJPE (mm): 10.25 -> 10.43 | Hurt |
| Use depth and point-cloud inputs rather than RGB-only input | IPNet · DexYCB AVG | MPJPE (mm): 14.04 -> 8.36 | Helped |
Closest papers by meaning
- Hand PointNet: 3D Hand Pose Estimation Using Point Sets (CVPR 2018 · similarity 0.94)
- Point-To-Pose Voting Based Hand Pose Estimation Using Residual Permutation Equivariant Layer (CVPR 2019 · similarity 0.92)
- 3D Convolutional Neural Networks for Efficient and Robust Hand Pose Estimation From Single Depth Images (CVPR 2017 · similarity 0.91)
- Depth-Based 3D Hand Pose Estimation: From Current Achievements to Future Goals (CVPR 2018 · similarity 0.91)
- Keypoint Fusion for RGB-D Based 3D Hand Pose Estimation (AAAI 2024 · similarity 0.88)
- HandDiff: 3D Hand Pose Estimation with Diffusion on Image-Point Cloud (CVPR 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.