Color Matters: Demosaicing-Guided Color Correlation Training for Generalizable AI-Generated Image Detection

TL;DR

This paper introduces a novel demosaicing-guided training framework that leverages intrinsic color correlations from camera pipelines to improve the detection of AI-generated images, especially on unseen generators.

Contribution

The proposed DCCT framework models color correlations induced by CFA and demosaicing, enhancing generalization in AI-generated image detection beyond prior methods.

Findings

Achieves state-of-the-art generalization on unseen generators.

Effectively models color correlation features for detection.

Significantly outperforms prior methods in robustness.

Abstract

As realistic AI-generated images threaten digital authenticity, we address the generalization failure of generative artifact-based detectors by exploiting the intrinsic properties of the camera imaging pipeline. Concretely, we investigate color correlations induced by the color filter array (CFA) and demosaicing, and propose a Demosaicing-guided Color Correlation Training (DCCT) framework for AI-generated image detection. By simulating the CFA sampling pattern, we decompose each color image into a single-channel input (as the condition) and the remaining two channels as the ground-truth targets (for prediction). A self-supervised U-Net is trained to model the conditional distribution of the missing channels from the given one, parameterized via a mixture of logistic functions. Our theoretical analysis reveals that DCCT targets a provable distributional difference in color-correlation features between photographic and AI-generated images. By leveraging these distinct features to construct a binary classifier, DCCT achieves state-of-the-art generalization and robustness, significantly outperforming prior methods across over 20 unseen generators.