SimLBR: Learning to Detect Fake Images by Learning to Detect Real Images

TL;DR

SimLBR is a novel fake image detection framework that learns a decision boundary around real images, significantly improving cross-generator generalization and efficiency, with a focus on robustness and reliability.

Contribution

Proposes SimLBR, a simple, efficient method using Latent Blending Regularization to improve fake image detection and robustness against distribution shifts.

Findings

Achieves up to +24.85% accuracy on Chameleon benchmark

Significantly improves cross-generator generalization

Training is orders of magnitude faster than existing methods

Abstract

The rapid advancement of generative models has made the detection of AI-generated images a critical challenge for both research and society. Recent works have shown that most state-of-the-art fake image detection methods overfit to their training data and catastrophically fail when evaluated on curated hard test sets with strong distribution shifts. In this work, we argue that it is more principled to learn a tight decision boundary around the real image distribution and treat the fake category as a sink class. To this end, we propose SimLBR, a simple and efficient framework for fake image detection using Latent Blending Regularization (LBR). Our method significantly improves cross-generator generalization, achieving up to +24.85\% accuracy and +69.62\% recall on the challenging Chameleon benchmark. SimLBR is also highly efficient, training orders of magnitude faster than existing approaches. Furthermore, we emphasize the need for reliability-oriented evaluation in fake image detection, introducing risk-adjusted metrics and worst-case estimates to better assess model robustness. All code and models will be released on HuggingFace and GitHub.