Research map / 3D vision / Human pose estimation
Semi-Supervised Unconstrained Head Pose Estimation in the Wild
Huayi Zhou, Fei Jiang, Jin Sha Yuan, Yong Rui, Hong Tao Lu, Kui Jia
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 aspect-ratio invariant cropping to FSA-Net (ablation) | FSA-Net · AFLW2000 | MAE: 5.07 -> 4.39 | Improved |
| add aspect-ratio invariant cropping to 6DRepNet (ablation) | 6DRepNet · AFLW2000 | MAE: 3.97 -> 3.49 | Improved |
| use dynamic entropy-based threshold instead of fixed threshold (ablation) | FisherMatch · AFLW2000 Setting1 20% labels | MAE | Improved |
| use CutMix instead of CutOut (ablation) | SemiUHPE · Setting1/3 | HPE error | Improved |
| combine CutOut and CutMix as T_CutOcc (ablation) | SemiUHPE · DAD-3DHeads | HPE error | Improved |
| add rotation consistency augmentation T_Rot (ablation) | SemiUHPE · DAD-3DHeads | HPE error | Improved |
| apply rotation consistency to 300W-LP (ablation) | SemiUHPE · 300W-LP | HPE error | No clear effect |
| use COCOHead rather than WiderFace as unlabeled set (ablation) | SemiUHPE · DAD-3DHeads val-set | angle error: 5.300 -> 4.800 | Improved |
| choose unsupervised loss weight 1.0 (ablation) | SemiUHPE · Setting3 | HPE error | No clear effect |
| SemiUHPE vs supervised with 2% labels | ResNet50 · AFLW2000 300W-LP 2% | MAE: 4.347 -> 3.956 | Improved |
| SemiUHPE vs FisherMatch at 50 labels per class | FisherMatch · Pascal3D+ | Acc@30deg: 75.7 -> 83.1 | Improved |
| add extra unlabeled data to SemiUHPE EffNetV2-S | SemiUHPE · DAD-3DHeads test-set | angle error: 4.636 -> 4.408 | Improved |
Closest papers by meaning
- Unsupervised Discovery of Facial Landmarks and Head Pose (CVPR 2025 · similarity 0.87)
- 6DoF Head Pose Estimation through Explicit Bidirectional Interaction with Face Geometry (ECCV 2024 · similarity 0.86)
- 3D Human Pose Estimation in Video With Temporal Convolutions and Semi-Supervised Training (CVPR 2019 · similarity 0.86)
- Learning Pose-Aware Models for Pose-Invariant Face Recognition in the Wild (TPAMI 2018 · similarity 0.85)
- Multi-Task Head Pose Estimation in-the-Wild (TPAMI 2020 · similarity 0.85)
- Category-Level 6D Object Pose Estimation in the Wild: A Semi-Supervised Learning Approach and A New Dataset (NeurIPS 2022 · similarity 0.85)
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.
