Beyond Artifacts: Real-Centric Envelope Modeling for Reliable AI-Generated Image Detection

TL;DR

This paper introduces Real-centric Envelope Modeling (REM), a novel approach that enhances the robustness of AI-generated image detection by modeling real image distributions rather than relying on generator artifacts, and demonstrates superior performance on diverse benchmarks.

Contribution

The paper proposes REM, a new paradigm for detection that focuses on real image distribution modeling, and introduces the RealChain benchmark for evaluating detection under real-world degradations.

Findings

REM achieves 7.5% higher accuracy than state-of-the-art methods.

REM maintains strong generalization on severely degraded images.

The RealChain benchmark covers diverse real-world degradations.

Abstract

The rapid progress of generative models has intensified the need for reliable and robust detection under real-world conditions. However, existing detectors often overfit to generator-specific artifacts and remain highly sensitive to real-world degradations. As generative architectures evolve and images undergo multi-round cross-platform sharing and post-processing (chain degradations), these artifact cues become obsolete and harder to detect. To address this, we propose Real-centric Envelope Modeling (REM), a new paradigm that shifts detection from learning generator artifacts to modeling the robust distribution of real images. REM introduces feature-level perturbations in self-reconstruction to generate near-real samples, and employs an envelope estimator with cross-domain consistency to learn a boundary enclosing the real image manifold. We further build RealChain, a comprehensive benchmark covering both open-source and commercial generators with simulated real-world degradation. Across eight benchmark evaluations, REM achieves an average improvement of 7.5% over state-of-the-art methods, and notably maintains exceptional generalization on the severely degraded RealChain benchmark, establishing a solid foundation for synthetic image detection under real-world conditions. The code and the RealChain benchmark will be made publicly available upon acceptance of the paper.