Research map / Image generation / Editing and style transfer
RSFNet: A White-Box Image Retouching Approach using Region-Specific Color Filters
Wenqi Ouyang, Yi Dong, Xiaoyang Kang, Peiran Ren, Xin Xu, Xuansong Xie
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 region-specific attention maps to filters (RSFNet-map vs RSFNet-global) (ablation) | RSFNet · FiveK 480p zeroed as shot | PSNR: 23.43 -> 24.64 | Improved |
| use parallel linear summation of filters instead of sequential cascade (ablation) | RSFNet · FiveK, same filters and backbone | PSNR / SSIM | Improved |
| pretrain mask generation first (PoolNet) then train arguments separately (ablation) | RSFNet-saliency · FiveK zeroed as shot | PSNR: 24.01 -> 23.26 | Hurt |
| concatenate off-the-shelf masks as network input (ablation) | RSFNet variants · FiveK zeroed as shot | PSNR: 24.17 -> 21.77 (semantic(5)) | Hurt |
| use more semantic masks (133 instead of 5) (ablation) | RSFNet-semantic · FiveK zeroed as shot | PSNR: 24.17 -> 23.75 | Hurt |
| decrease backbone downscaling from x8 to x4 (resnet18) (ablation) | RSFNet · FiveK zeroed as shot | PSNR: 24.65 -> 24.83 (runtime 9.98 -> 12.38 ms) | Improved |
| controlled-shape masks (palette, saliency, semantic) versus free region maps | RSFNet · FiveK 480p zeroed with C's | PSNR: 25.49 -> 25.01 / 24.78 / 24.76 | Hurt |
| RSFNet-map versus 3D-LUT+AdaInt on PPR10K | RSFNet-map · PPR10K expert a | PSNR: 25.98 -> 25.58 | Mixed |
Closest papers by meaning
- Conditional Sequential Modulation for Efficient Global Image Retouching (ECCV 2020 · similarity 0.88)
- ChromaFusionNet (CFNet): Natural Fusion of Fine-Grained Color Editing (AAAI 2024 · similarity 0.86)
- Close Imitation of Expert Retouching for Black-and-White Photography (CVPR 2024 · similarity 0.86)
- PerTouch: VLM-Driven Agent for Personalized and Semantic Image Retouching (AAAI 2026 · similarity 0.86)
- DiffRetouch: Using Diffusion to Retouch on the Shoulder of Experts (AAAI 2025 · similarity 0.86)
- RetouchFormer: Semi-supervised High-Quality Face Retouching Transformer with Prior-Based Selective Self-Attention (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.
