Research map / 3D vision / Human pose estimation
Feature Boosting Network For 3D Pose Estimation
Jun Liu, Henghui Ding, Amir Shahroudy, Ling‐Yu Duan, Xudong Jiang, Gang Wang, Alex Chichung Kot
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 LSTD feature boosting (FB) to 3D Pose Net | 3D Pose Net · 3DHandPose | PCK@20: 85.7% -> 87.7% | Improved |
| add context consistency gate (CCG) to FB (ablation) | 3D Pose Net with FB · 3DHandPose | PCK@20: 87.7% -> 89.5% | Improved |
| add FB to 3D Pose Net | 3D Pose Net · Human3.6M | PCK@50: 48.1% -> 49.8% | Improved |
| add CCG to FB (ablation) | 3D Pose Net with FB · Human3.6M | PCK@50: 49.8% -> 51.5% | Improved |
| add FB+ for cross-dataset evaluation | 3D Pose Net · MPI-INF-3DHP | PCK: 66.2% -> 69.6% | Improved |
| use graphical (physical + symmetrical) instead of physical-only dependency (ablation) | LSTD module · 3DHandPose | PCK@20: 87.5% -> 89.0% | Improved |
| add backward pass (bi-directional) to graphical dependency (ablation) | LSTD module · 3DHandPose | PCK@20: 89.0% -> 89.5% | Improved |
| use graph dependency instead of simple sequence chain (ablation) | LSTD module · 3DHandPose | PCK@20: 86.1% -> 89.5% | Improved |
| stack two sub-networks instead of one (ablation) | Feature boosting network · 3DHandPose | PCK@20: 87.4% -> 89.5% | Improved |
| add extra dependency links (max linked joints 4 -> 7) (ablation) | FB+ · Human3.6M | PCK@50: 51.5% -> 51.7% | No clear effect |
| add extra dependency links (max linked joints 4 -> 8) (ablation) | FB+ · 3DHandPose | PCK@20: 89.5% -> 89.5% | No clear effect |
| use ConvLSTM instead of ConvRNN (ablation) | 3D Pose Net with FB · 3DHandPose | PCK@20: 87.0% -> 87.7% | Improved |
Closest papers by meaning
- Real-Time 3D Hand Pose Estimation with 3D Convolutional Neural Networks (TPAMI 2018 · similarity 0.90)
- Exploiting Spatial-Temporal Relationships for 3D Pose Estimation via Graph Convolutional Networks (ICCV 2019 · similarity 0.90)
- 3D Hand Shape and Pose Estimation From a Single RGB Image (CVPR 2019 · similarity 0.90)
- 3D Convolutional Neural Networks for Efficient and Robust Hand Pose Estimation From Single Depth Images (CVPR 2017 · similarity 0.90)
- Learning to Estimate 3D Hand Pose From Single RGB Images (ICCV 2017 · similarity 0.90)
- Recurrent 3D Hand Pose Estimation Using Cascaded Pose-Guided 3D Alignments (TPAMI 2022 · similarity 0.89)
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.
