Research map / 3D vision / Face and shape
Face Alignment in Full Pose Range: A 3D Total Solution
Xiangyu Zhu, Xiaoming Liu, Zhen Lei, Stan Ziqing Li
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 |
|---|---|---|---|
| train on face-profiled 300W-LP instead of 300W | RCPR · AFLW (21 pts) mean | NME (%): 11.94 -> 7.85 | Improved |
| train on face-profiled 300W-LP instead of 300W | ESR · AFLW (21 pts) mean | NME (%): 12.07 -> 8.24 | Improved |
| train on face-profiled 300W-LP instead of 300W | SDM · AFLW (21 pts) mean | NME (%): 9.19 -> 6.55 | Improved |
| train on face-profiled 300W-LP instead of 300W | SDM · AFLW2000-3D yaw 60-90 | NME (%): 17.48 -> 9.76 | Improved |
| fit dense 3DMM with cascaded CNN (3DDFA) instead of SDM on 300W-LP | 3DDFA · AFLW (21 pts) mean | NME (%): 6.55 -> 5.60 | Improved |
| add SDM landmark refinement after 3DDFA | 3DDFA · AFLW (21 pts) mean | NME (%): 5.60 -> 5.32 | Improved |
| add SDM landmark refinement after 3DDFA | 3DDFA · AFLW2000-3D mean | NME (%): 5.42 -> 4.94 | Improved |
| add SDM landmark refinement after 3DDFA | 3DDFA · 300W Full set | NME (%): 7.01 -> 6.31 | Improved |
| add initialization regeneration to the cascade (ablation) | 3DDFA · 300W-LP held-out test error | NME (%) | Improved |
| use WPDC cost instead of PDC or VDC (ablation) | 3DDFA · 300W-LP testing error | NME (%) | Improved |
| use 3DDFA+SDM on 300W medium poses vs CFSS | 3DDFA+SDM · 300W Common set | NME (%): 4.73 -> 5.53 | Hurt |
| use 3DDFA+SDM instead of SDM on challenging set | 3DDFA+SDM · 300W Challenging set | NME (%): 15.40 -> 9.56 | Improved |
Closest papers by meaning
- Face Alignment Across Large Poses: A 3D Solution (CVPR 2016 · similarity 0.99)
- How Far Are We From Solving the 2D & 3D Face Alignment Problem? (And a Dataset of 230,000 3D Facial Landmarks) (ICCV 2017 · similarity 0.90)
- Large-Pose Face Alignment via CNN-Based Dense 3D Model Fitting (CVPR 2016 · similarity 0.90)
- Faster Than Real-Time Facial Alignment: A 3D Spatial Transformer Network Approach in Unconstrained Poses (ICCV 2017 · similarity 0.89)
- Large Pose 3D Face Reconstruction From a Single Image via Direct Volumetric CNN Regression (ICCV 2017 · similarity 0.89)
- FaceLift: Semi-supervised 3D Facial Landmark Localization (CVPR 2024 · similarity 0.86)
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.
