Research map / Image generation / Editing and style transfer
Leveraging Color Naming for Image Enhancement
David Serrano-Lozano, Luis Herranz, Michael S. Brown, Javier Vázquez-Corral
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 standardization backbone (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 24.30 -> 25.75 | Improved |
| add color naming decomposition (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 25.13 -> 25.75 | Improved |
| add transformer fusion block (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 25.41 -> 25.75 | Improved |
| replace conference NamedCurves with NamedCurves+ (transformer fusion) | NamedCurves · MIT-Adobe-5K retouching | PSNR: 25.59 -> 25.75 | Improved |
| transformer fusion also lowers inference time | NamedCurves+ · MIT-Adobe-5K retouching | inference time (ms): 26 -> 19 | Improved |
| use perceptual color-naming model (Van de Weijer, 6 colors) instead of random GMM (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 24.12 -> 25.75 | Improved |
| use 6 color names instead of 11 (ablation) | NamedCurves+ with Van de Weijer · MIT-Adobe-5K retouching | PSNR: 25.49 -> 25.75 | Improved |
| use 11 control points instead of 7 (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 25.70 -> 25.75 | Improved |
| use 11 control points instead of 16 (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 25.59 -> 25.75 | Improved |
| use loss weight alpha 0.5 instead of 0 (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 25.41 -> 25.75 | Improved |
| use loss weight alpha 0.5 instead of 1 (ablation) | NamedCurves+ · MIT-Adobe-5K retouching | PSNR: 25.58 -> 25.75 | Improved |
| extend NamedCurves+ to exposure correction | NamedCurves+ · SICE exposure correction (average) | PSNR: 21.92 -> 22.08 | Improved |
Closest papers by meaning
- Learned Image Enhancement via Color Naming (ECCV 2024 · similarity 0.94)
- Recolour What Matters: Region-Aware Colour Editing via Token-Level Diffusion (ECCV 2026 · similarity 0.88)
- Deep Edge-Aware Interactive Colorization Against Color-Bleeding Effects (ICCV 2021 · similarity 0.87)
- Automatic Controllable Colorization via Imagination (CVPR 2024 · similarity 0.86)
- ChromaFusionNet (CFNet): Natural Fusion of Fine-Grained Color Editing (AAAI 2024 · similarity 0.85)
- Close Imitation of Expert Retouching for Black-and-White Photography (CVPR 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.
