Research map / 3D vision / Human pose estimation
Learning to Estimate 3D Hand Pose From Single RGB Images
Christian Zimmermann, Thomas Brox
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 |
|---|---|---|---|
| predict in canonical frame plus viewpoint instead of direct 3D lifting (ablation) | 3D lifting network · R-val | median end point error (mm): 20.9 -> 18.8 | Improved |
| predict in canonical frame plus viewpoint instead of bottleneck layer (ablation) | 3D lifting network · R-val | median end point error (mm): 21.9 -> 18.8 | Improved |
| predict in canonical frame instead of local kinematic angles (ablation) | 3D lifting network · R-val | median end point error (mm): 39.1 -> 18.8 | Improved |
| predict in canonical frame instead of nearest-neighbour retrieval (ablation) | 3D lifting network · R-val | median end point error (mm): 26.9 -> 18.8 | Improved |
| use bottleneck layer instead of direct regression (ablation) | 3D lifting network · R-val | median end point error (mm): 20.9 -> 21.9 | No clear effect |
| replace GT hand crops with HandSegNet crops | PoseNet · S-val | AUC: 0.817 -> 0.762 | Hurt |
| replace GT hand crops with HandSegNet crops | PoseNet · R-val | mean end point error (px): 9.135 -> 18.741 | Hurt |
| train on synthetic R-train plus S-train instead of S-train alone (ablation) | PoseNet · Dexter | PCK | Improved |
| use 3D hand pose features for gesture recognition versus prior method | GestureNet · RWTH German Fingerspelling | word error rate (%): 36.56 -> 33.2 | Improved |
| feed score maps instead of coordinates to the lifting network (ablation) | PosePrior · S-val | PCK | Improved |
Closest papers by meaning
- Real-Time 3D Hand Pose Estimation with 3D Convolutional Neural Networks (TPAMI 2018 · similarity 0.94)
- 3D Hand Shape and Pose Estimation From a Single RGB Image (CVPR 2019 · similarity 0.94)
- 3D Convolutional Neural Networks for Efficient and Robust Hand Pose Estimation From Single Depth Images (CVPR 2017 · similarity 0.94)
- DeepHand: Robust Hand Pose Estimation by Completing a Matrix Imputed With Deep Features (CVPR 2016 · similarity 0.92)
- Depth-Based 3D Hand Pose Estimation: From Current Achievements to Future Goals (CVPR 2018 · similarity 0.92)
- 3D Hand Pose Estimation Using Synthetic Data and Weakly Labeled RGB Images (TPAMI 2020 · similarity 0.92)
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.
