NamedCurves: Learned Image Enhancement via Color Naming
David Serrano-Lozano, Luis Herranz, Michael S. Brown, Javier Vazquez-Corral
Computer Vision Center, Universitat Autònoma de Barcelona, Universidad Autónoma de Madrid, York University
NamedCurves is a novel image enhancement method that decomposes an image into a small set of named colors and enhances them using a learned set of tone curves. This decomposition improves the model’s interactivity, allowing users to manipulate specific colors in the image by adjusting the corresponding tone curves.
- Automatic Enhancement. Upload an image and click “Run” to apply automatic enhancement based on the learned tone curves.
- Interactive Adjustments. Fine-tune the enhancement by modifying the tone curves for each named color clicking "Recompute". This lets you manipulate specific colors without affecting others in the image. Each color has a set of control points that define the tone curves. Adjust the R, G, and B values for each control point with the sliders below the curve visualization.
Explore the tool and create customized, color-aware image enhancements effortlessly!
@article{serrano2024namedcurves,
author = {Serrano-Lozano, David and Herranz, Luis and Brown, Michael S. and Vazquez-Corral, Javier},
title = {NamedCurves: Learned Image Enhancement via Color Naming},
journal = {ECCV},
year = {2024}}
Examples
0 1
0 1
0 1
0 1
0 1
0 1
0 1
0 1
0 1
0 1
0 1