Research map / Image generation / Editing and style transfer
Learning Flow-based Feature Warping for Face Frontalization with Illumination Inconsistent Supervision
Yuxiang Wei, Ming Liu, Haolin Wang, Ruifeng Zhu, Guosheng Hu, Wangmeng Zuo
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 Illumination Preserving Module (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 98.00 -> 98.04 | Improved |
| add illumination preserving pathway (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.62 -> 98.04 | Improved |
| add illumination adaption pathway (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.92 -> 98.04 | Improved |
| add Warp Attention Module (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 95.81 -> 98.04 | Improved |
| add flipped warped feature in WAM (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.69 -> 98.04 | Improved |
| add feature warping in WAM (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 96.54 -> 98.04 | Improved |
| add feature attention in WAM (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.94 -> 98.04 | Improved |
| add adversarial loss (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.93 -> 98.04 | Improved |
| add identity preserving loss (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 96.00 -> 98.04 | Improved |
| add pixel-wise loss (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.94 -> 98.04 | Improved |
| add perceptual loss (ablation) | FFWM · Multi-PIE Setting 2 (average) | Rank-1 accuracy (%): 97.96 -> 98.04 | Improved |
| add Warp Attention Module (ablation) | FFWM · Multi-PIE Setting 2 (+-90 deg) | Rank-1 accuracy (%): 86.23 -> 93.17 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DA-GAN — DA-GAN: Instance-Level Image Translation by Deep Attention Generative Adversarial Networks (CVPR 2018) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Towards High Fidelity Face Frontalization in the Wild (IJCV 2019 · similarity 0.85)
- Learning a High Fidelity Pose Invariant Model for High-resolution Face Frontalization (NeurIPS 2018 · similarity 0.85)
- Neural Implicit Morphing of Face Images (CVPR 2024 · similarity 0.84)
- ControlFace: Harnessing Facial Parametric Control for Face Rigging (CVPR 2025 · similarity 0.82)
- Generalizable Face Landmarking Guided by Conditional Face Warping (CVPR 2024 · similarity 0.81)
- Comprehensive Relighting: Generalizable and Consistent Monocular Human Relighting and Harmonization (CVPR 2025 · similarity 0.81)
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.
