Research map / 3D vision / Human pose estimation
Depth-Based 3D Hand Pose Estimation: From Current Achievements to Future Goals
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 |
|---|---|---|---|
| Use a 3D CNN and volumetric representation instead of a 2D CNN | Hierarchical hand pose estimation · Single frame pose estimation | success rate | Helped |
| Use detection-based estimation rather than regression-based estimation | Single frame pose estimation · Single frame pose estimation | success rate | Helped |
| Add hierarchical finger and joint-type constraints | rvhand · Single frame pose estimation with occluded joints | success rate | Helped |
| Add structural constraints in the loss function rather than within CNN layers | Structured hand pose estimation · Single frame pose estimation | success rate | Helped |
| Represent pose with phalanges and constrain their long-range interactions | strawberryfg · Single frame pose estimation | success rate | Helped |
| Use multiple cascaded stages to update pose estimates | THU_VCLab versus rvhand with REN backbone · Single frame pose estimation | success rate | Helped |
| Use a depth-prediction stage | RCN-3D · Single frame pose estimation at view points [150,180] degrees | mean error (mm): 2-3 mm | Helped |
| Use motion estimation with single-frame hand pose estimation | RCN-3D_track · Hand pose tracking | average error (mm): 10.5 vs 12.6 and 13.7 | Helped |
| Use per-frame hand detection | NAIST_RV_track · Hand pose tracking | tracking success rate | Helped |
| Use CNN-based hand-object segmentation rather than image-processing object removal | Hand-object pose estimation · Hand-object interaction, visible and occluded joints | mean error | Helped |
| Train discriminative methods on seen and unseen hand shapes | Single frame pose estimation · Unseen hand shapes | mean error (mm): 7.0 -> 12.2 | Hurt |
| Increase the number of visible joints | Single frame pose estimation · Frames grouped by number of occluded joints | mean error | Helped |
Closest papers by meaning
- 3D Convolutional Neural Networks for Efficient and Robust Hand Pose Estimation From Single Depth Images (CVPR 2017 · similarity 0.93)
- Robust 3D Hand Pose Estimation in Single Depth Images: From Single-View CNN to Multi-View CNNs (CVPR 2016 · similarity 0.93)
- Dense 3D Regression for Hand Pose Estimation (CVPR 2018 · similarity 0.93)
- DeepHand: Robust Hand Pose Estimation by Completing a Matrix Imputed With Deep Features (CVPR 2016 · similarity 0.92)
- Efficiently Creating 3D Training Data for Fine Hand Pose Estimation (CVPR 2016 · similarity 0.92)
- Keypoint Fusion for RGB-D Based 3D Hand Pose Estimation (AAAI 2024 · similarity 0.90)
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.