Semantic-Aware Reconstruction Error for Detecting AI-Generated Images

TL;DR

This paper introduces Semantic-Aware Reconstruction Error (SARE), a novel method for detecting AI-generated images by measuring semantic shifts during caption-guided reconstruction, leading to improved robustness across diverse models.

Contribution

The paper proposes SARE, a new semantic-aware feature for fake image detection, and a fusion module with cross-attention to enhance generalization to unseen generative models.

Findings

SARE outperforms existing methods on GenImage and ForenSynths benchmarks.

Semantic shifts are effective indicators for distinguishing real and fake images.

Caption-guided reconstruction improves detection robustness.

Abstract

Recently, AI-generated image detection has gained increasing attention, as the rapid advancement of image generation technologies has raised serious concerns about their potential misuse. While existing detection methods have achieved promising results, their performance often degrades significantly when facing fake images from unseen, out-of-distribution (OOD) generative models, since they primarily rely on model-specific artifacts and thus overfit to the models used for training. To address this limitation, we propose a novel representation, namely Semantic-Aware Reconstruction Error (SARE), that measures the semantic difference between an image and its caption-guided reconstruction. The key hypothesis behind SARE is that real images, whose captions often fail to fully capture their complex visual content, may undergo noticeable semantic shifts during the caption-guided reconstruction process. In contrast, fake images, which closely align with their captions, show minimal semantic changes. By quantifying these semantic shifts, SARE provides a robust and discriminative feature for detecting fake images across diverse generative models. Additionally, we introduce a fusion module that integrates SARE into the backbone detector via a cross-attention mechanism. Image features attend to semantic representations extracted from SARE, enabling the model to adaptively leverage semantic information. Experimental results demonstrate that the proposed method achieves strong generalization, outperforming existing baselines on benchmarks including GenImage and ForenSynths. We further validate the effectiveness of caption guidance through a detailed analysis of semantic shifts, confirming its ability to enhance detection robustness.