Research map / Image generation / Editing and style transfer
RefSTAR: Blind Face Image Restoration with Reference Selection, Transfer, and Reconstruction
Zhicun Yin, Junjie Chen, Ming Liu, Zhixin Wang, Fan Li, Renjing Pei, Xiaoming Li, Rynson W. H. Lau, 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 |
|---|---|---|---|
| use RefSel mask instead of all-zeros mask (ablation) | RefSTAR · Celeb-Ref-Test | ID_GT: 69.24 -> 78.04 | Improved |
| use RefSel mask instead of all-ones mask (ablation) | RefSTAR · Celeb-Ref-Test | ID_GT: 70.23 -> 78.04 | Improved |
| use RefSel mask instead of all-ones mask (ablation) | RefSTAR · Celeb-Ref-Test | ID_Ref: 69.34 -> 64.68 | Hurt |
| use RefSel mask instead of all-zeros mask (ablation) | RefSTAR · Celeb-Ref-Test | FID: 25.01 -> 22.21 | Improved |
| add mask-compatible cycle-consistency loss (ablation) | RefSTAR · Celeb-Ref-Test | ID_Ref: 43.64 -> 64.68 | Improved |
| add mask-compatible cycle-consistency loss (ablation) | RefSTAR · Celeb-Ref-Test | ID_GT: 67.21 -> 78.04 | Improved |
| add decoupled cross-attention (ablation) | RefSTAR · Celeb-Ref-Test | ID_Ref: 47.22 -> 64.68 | Improved |
| add decoupled cross-attention (ablation) | RefSTAR · Celeb-Ref-Test | FID: 25.29 -> 22.21 | Improved |
| add RefSTAR reference modules vs best baseline | RefSTAR · Celeb-Ref-Test | PSNR: 23.99 vs 24.34 (DMDNet) | No clear effect |
| RefSTAR vs RefLDM on real-world | RefSTAR · RealRef60 | ID_Ref: 47.33 -> 55.73 | 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.
- RefLDM — ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration (NeurIPS 2024) · better on 6 of 6 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
- ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration (NeurIPS 2024 · similarity 0.88)
- FaceMe: Robust Blind Face Restoration with Personal Identification (AAAI 2025 · similarity 0.88)
- RestoreFormer++: Towards Real-World Blind Face Restoration From Undegraded Key-Value Pairs (TPAMI 2023 · similarity 0.87)
- Overcoming False Illusions in Real-World Face Restoration with Multi-Modal Guided Diffusion Model (ICLR 2025 · similarity 0.85)
- Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery (AAAI 2024 · similarity 0.84)
- Low-Light Face Super-resolution via Illumination, Structure, and Texture Associated Representation (AAAI 2024 · similarity 0.84)
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.
