Chroma Clues: Leveraging Color Statistics to Detect Synthetic Images
2026-06-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors studied how AI-made images often have unusual color patterns because the way these AI systems are trained focuses less on color details. They created six special color changes and also a way to learn the best color adjustments to spot fake images. Using these color tweaks, a simple system can tell real from fake images with over 93% accuracy and still work well even if the images are changed a bit. Their method also helps visualize where fake colors show up and improves identifying the type of AI used to create images.
AI image generationimage forensicsLPIPS losschrominanceluminancecolor statisticsimage classificationpost-processingmulticlass attributionvisual noise patterns
Authors
Lea Uhlenbrock, Davide Cozzolino, Christian Riess
Abstract
The evolution and dissemination of AI-synthesized images is occurring at an unprecedented rate. Image generators are making rapid progress in their goal of perfectly imitating natural images, which also challenges image forensics. In this work, we exploit an underexplored cue in current generative models, namely their weakness to imitate color statistics of natural images. We first show that the LPIPS loss used for training image generators is less sensitive to chrominance than to luminance, which may lead to statistical discrepancies in the colors of synthetic images. Building on this observation, we then introduce six hand-crafted color transformations and a method to learn a task-optimized color transform to statistically expose generated images. These transformations can be used in various ways. First, we define color-sensitive features at pixel-level or patch-level. A simple, interpretable classifier achieves with these features an average generalization accuracy of 93.27% and strong robustness against six types of post-processing. Second, we demonstrate that the transformations exhibit characteristic visual noise patterns in natural and synthetic image areas, which enables an intuitive visual image evaluation. Third, we demonstrate that the transforms can enhance color patterns in generated images for improved multiclass attribution.