Research map / 3D vision / Human pose estimation
SemiHand: Semi-Supervised Hand Pose Estimation With Consistency
Linlin Yang, Shicheng Chen, Angela Yao
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 view consistency loss to the synthetic-pretrained baseline (ablation) | SemiHand baseline · STB test, mean EPE (mm) | mean EPE: 23.83 -> 21.03 | Improved |
| add cross-modal consistency loss to the baseline (ablation) | SemiHand baseline · STB test, mean EPE (mm) | mean EPE: 23.83 -> 20.92 | Improved |
| add both consistency losses (vc+cc) to the baseline (ablation) | SemiHand baseline · STB test, mean EPE (mm) | mean EPE: 23.83 -> 19.93 | Improved |
| add pseudo-labelling with label correction and sample selection to the baseline (ablation) | SemiHand baseline · STB test, mean EPE (mm) | mean EPE: 7.52 mm improvement (23.83 -> 16.31) | Improved |
| combine consistency training and pseudo-labelling (full SemiHand) (ablation) | SemiHand baseline · STB test, mean EPE (mm) | mean EPE: 9.23mm improvement (23.83 -> 14.60) | Improved |
| add pseudo-labelling on top of both consistency losses (ablation) | SemiHand · STB test, mean EPE (mm) | mean EPE: 19.93 -> 14.60 | Improved |
| fine-tune on the unlabelled STB test images instead of train images | SemiHand · STB test, mean EPE (mm) | mean EPE: 14.60 -> 14.51 | Improved |
| fine-tune on unlabelled STB train and test together | SemiHand · STB test, mean EPE (mm) | mean EPE: 14.60 -> 13.95 | Improved |
| increase the percentage of unlabelled STB training data | SemiHand · STB test, mean EPE (mm) | mean EPE: 17.31 -> 14.60 | Improved |
| semi-supervised fine-tuning on unlabelled real images across datasets | SemiHand baseline · Dexter+Object, mean EPE | mean EPE: decrease of up to 22.4 mm on DO | Improved |
| semi-supervised fine-tuning, AUC on STB | SemiHand baseline · STB AUC | AUC: 0.774 -> 0.927 | Improved |
| semi-supervised fine-tuning, AUC on Dexter+Object | SemiHand baseline · Dexter+Object AUC | AUC: 0.546 -> 0.747 | Improved |
Closest papers by meaning
- Semi-Supervised 3D Hand-Object Poses Estimation With Interactions in Time (CVPR 2021 · similarity 0.92)
- Efficiently Creating 3D Training Data for Fine Hand Pose Estimation (CVPR 2016 · similarity 0.90)
- Cross-Domain 3D Hand Pose Estimation With Dual Modalities (CVPR 2023 · similarity 0.90)
- Weakly-supervised 3D Hand Pose Estimation from Monocular RGB Images (ECCV 2018 · similarity 0.90)
- 3D Hand Pose Estimation Using Synthetic Data and Weakly Labeled RGB Images (TPAMI 2020 · similarity 0.89)
- SiMHand: Mining Similar Hands for Large-Scale 3D Hand Pose Pre-training (ICLR 2025 · similarity 0.87)
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.
