Research map / Image generation / Image restoration
Enhanced Blind Face Restoration With Multi-Exemplar Images and Adaptive Spatial Feature Fusion
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 |
|---|---|---|---|
| Use ASFFNet with multi-exemplar guidance, MLS alignment, AdaIN, and adaptive progressive fusion | Blind face restoration · VGGFace2 ×4 SR | PSNR: 28.07 vs 27.66 | Helped |
| Use ASFFNet with multi-exemplar guidance, MLS alignment, AdaIN, and adaptive progressive fusion | Blind face restoration · VGGFace2 ×8 SR | PSNR: 24.34 vs 23.87 | Helped |
| Increase guidance exemplar count from 1 to 10 (ablation) | ASFFNet · VGGFace2 ×4 SR | PSNR: 27.99 -> 28.07 | Helped |
| Increase guidance exemplar count from 1 to 10 (ablation) | ASFFNet · VGGFace2 ×8 SR | PSNR: 24.19 -> 24.34 | Helped |
| Use four adaptive spatial feature-fusion blocks instead of one (ablation) | ASFFNet · VGGFace2 ×4 SR | PSNR: 27.59 -> 28.07 | Helped |
| Use eight rather than four adaptive spatial feature-fusion blocks (ablation) | ASFFNet · VGGFace2 ×4 SR | PSNR: 28.07 -> 28.07 | No clear effect |
| Replace adaptive fusion with concatenation (ablation) | ASFFNet · VGGFace2 ×4 SR | PSNR: 28.07 -> 27.56 | Hurt |
| Add attention masks to the fusion blocks (ablation) | ASFFNet · VGGFace2 ×4 SR | PSNR: 27.83 -> 28.07 | Helped |
| Add AdaIN illumination translation (ablation) | ASFFNet · VGGFace2 ×8 SR | PSNR: 24.15 -> 24.34 | Helped |
| Add MLS spatial alignment (ablation) | ASFFNet · VGGFace2 ×8 SR | PSNR: 23.12 -> 24.34 | Helped |
| Train the guidance feature extractor rather than keeping it fixed (ablation) | ASFFNet · VGGFace2 ×8 SR | PSNR: 23.47 -> 24.34 | Helped |
Closest papers by meaning
- Blind Face Restoration under Extreme Conditions: Leveraging 3D-2D Prior Fusion for Superior Structural and Texture Recovery (AAAI 2024 · similarity 0.87)
- Overcoming False Illusions in Real-World Face Restoration with Multi-Modal Guided Diffusion Model (ICLR 2025 · similarity 0.85)
- Self-Supervised Selective-Guided Diffusion Model for Old-Photo Face Restoration (NeurIPS 2025 · similarity 0.85)
- ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration (NeurIPS 2024 · similarity 0.85)
- $InterLCM$: Low-Quality Images as Intermediate States of Latent Consistency Models for Effective Blind Face Restoration (ICLR 2025 · similarity 0.85)
- FaceMe: Robust Blind Face Restoration with Personal Identification (AAAI 2025 · 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.