Research map / Image generation / Editing and style transfer
Blind Face Restoration via Deep Multi-scale Component Dictionaries
Xiaoming Li, Chaofeng Chen, Shangchen Zhou, Xianhui Lin, Wangmeng Zuo, Lei Zhang
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 |
|---|---|---|---|
| increase dictionary cluster number K from 16 to 64 (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 26.79 -> 27.15 | Improved |
| increase dictionary cluster number K from 64 to 128 (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 27.15 -> 27.43 | Improved |
| increase dictionary cluster number K from 128 to 256 (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 27.43 -> 27.54 | Improved |
| increase dictionary cluster number K from 256 to 512 (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 27.54 -> 27.55 | No clear effect |
| increase K from 16 to 256 (ablation) | DFDNet256 · VggFace2 x4 SR | LPIPS: .144 -> .114 | Improved |
| add progressive DFT blocks instead of SFT only (0DFT) (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 25.30 -> 27.54 | Improved |
| use four DFT blocks instead of two (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 26.43 -> 27.54 | Improved |
| add CAdaIN in DFT blocks (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 25.47 -> 27.54 | Improved |
| add CAdaIN in DFT blocks (ablation) | DFDNet256 · VggFace2 x4 SR | LPIPS: .190 -> .114 | Improved |
| add confidence score in DFT blocks (ablation) | DFDNet256 · VggFace2 x4 SR | PSNR: 27.23 -> 27.54 | Improved |
| add confidence score in DFT blocks (ablation) | DFDNet256 · VggFace2 x4 SR | LPIPS: .129 -> .114 | Improved |
| add multi-scale component dictionaries (DFDNet256 vs GFRNet reference-based) | DFDNet256 · VggFace2 x4 SR | PSNR: 27.13 -> 27.54 | Improved |
Closest papers by meaning
- Learning Dual Memory Dictionaries for Blind Face Restoration (TPAMI 2022 · similarity 0.93)
- RestoreFormer++: Towards Real-World Blind Face Restoration From Undegraded Key-Value Pairs (TPAMI 2023 · similarity 0.90)
- A Natural Language Guided Approach for Blind Face Restoration: Methodology and Dataset (TPAMI 2026 · similarity 0.89)
- RestoreFormer: High-Quality Blind Face Restoration From Undegraded Key-Value Pairs (CVPR 2022 · similarity 0.88)
- FaceMe: Robust Blind Face Restoration with Personal Identification (AAAI 2025 · similarity 0.87)
- OSDFace: One-Step Diffusion Model for Face Restoration (CVPR 2025 · 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.
