CanonCGT: Reference-Based Color Grading via Canonical Pivot Representation
2026-06-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors propose CanonCGT, a new method for color grading photos by copying the tone and lighting style from a reference image while keeping colors natural and stable. Their approach uses a two-step process: first, it removes any original color bias to create a neutral version of the photo, and then it adjusts this version to match the reference style. They also introduce a special training method that mixes learning from preset examples and refining with unpaired photos. Their method produces more consistent and realistic results than previous techniques.
color gradingtonal moodphotorealistic renderingtone mappingcolor harmonyself-supervised learningstyle transferdual-phase trainingimage processingreference-based color grading
Authors
Jinwon Ko, Keunsoo Ko, Chang-Su Kim
Abstract
Reference-based color grading aims to reproduce the tonal mood and lighting of a reference while preserving color harmony and scene structure. Existing photorealistic and filter-based methods often produce unstable tone mappings -- over-shifting or inconsistently retaining colors -- leading to unnatural results. We propose CanonCGT, a two-stage framework built on a canonical pivot -- a style-neutral intermediate representation for stable color mapping. The first stage canonicalizes the input by removing intrinsic tonal bias, and the second color-grades it to match the reference style. A dual-phase training scheme, DP-CGT, combines supervised preset learning with self-supervised refinement on unpaired photographs. CanonCGT delivers photorealistic and tonally consistent results across diverse datasets, surpassing state-of-the-art methods in stability and visual fidelity. Our codes are available at \href{https://github.com/Jinwon-Ko/CanonCGT}{https://github.com/Jinwon-Ko/CanonCGT}