Research map / 3D vision / Face and shape
Learning To Restore 3D Face From In-the-Wild Degraded Images
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the StyleGAN prior (ablation) | L2R · BFM | SIDE (×10−2): 0.802±0.168 -> 0.710±0.139 | Helped |
| Add multi-style codes (ablation) | L2R · BFM | SIDE (×10−2): 0.735±0.124 -> 0.710±0.139 | Helped |
| Add spatial guidance (ablation) | L2R · BFM | SIDE (×10−2): 0.732±0.138 -> 0.710±0.139 | Helped |
| Add the prior decoder (ablation) | L2R · BFM | SIDE (×10−2): 0.728±0.129 -> 0.710±0.139 | Helped |
| Use the restored image as input to geometry refinement (ablation) | L2R · BFM | SIDE (×10−2): 0.776±0.178 -> 0.710±0.139 | Helped |
| Add the displacement map (ablation) | L2R · BFM | SIDE (×10−2): 0.725±0.178 -> 0.710±0.139 | Helped |
| Add the cross-view loss (ablation) | L2R · BFM | SIDE (×10−2): 0.727±0.140 -> 0.710±0.139 | Helped |
| Add the geometry refining network (ablation) | L2R · Texture test set | Cosine-Similarity: 0.703 -> 0.725 | Helped |
| Use DFDNet preprocessing instead of the StyleGAN prior (ablation) | L2R · BFM | SIDE (×10−2): 0.729±0.145 vs 0.710±0.139 | Hurt |
Closest papers by meaning
- Towards High-Fidelity Face Self-Occlusion Recovery via Multi-View Residual-Based GAN Inversion (AAAI 2022 · similarity 0.89)
- Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery (AAAI 2024 · similarity 0.89)
- Towards High-Fidelity 3D Face Reconstruction From In-the-Wild Images Using Graph Convolutional Networks (CVPR 2020 · similarity 0.89)
- GANFIT: Generative Adversarial Network Fitting for High Fidelity 3D Face Reconstruction (CVPR 2019 · similarity 0.89)
- FreeUV: Ground-Truth-Free Realistic Facial UV Texture Recovery via Cross-Assembly Inference Strategy (CVPR 2025 · similarity 0.87)
- In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (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.