Abstract
With generative models becoming increasingly sophisticated and diverse, detecting AI-generated images has become increasingly challenging. While existing AI-genereted Image detectors achieve promising performance on in-distribution generated images, their generalization to unseen generative models remains limited. This limitation is largely attributed to their reliance on generation-specific artifacts, such as stylistic priors and compression patterns. To address these limitations, we propose GAMMA, a novel training framework designed to reduce domain bias and enhance semantic alignment. GAMMA introduces diverse manipulation strategies, such as inpainting-based manipulation and semantics-preserving perturbations, to ensure consistency between manipulated and authentic content. We employ multi-task supervision with dual segmentation heads and a classification head, enabling pixel-level source attribution across diverse generative domains. In addition, a reverse cross-attention mechanism is introduced to allow the segmentation heads to guide and correct biased representations in the classification branch. Our method achieves state-of-the-art generalization performance on the GenImage benchmark, imporving accuracy by 5.8%, but also maintains strong robustness on newly released generative model such as GPT-4o.
Abstract (translated)
随着生成模型变得越来越复杂和多样化，检测AI生成的图像也变得更加具有挑战性。现有的针对AI生成图像的检测器在处理同分布（in-distribution）生成图像时表现出色，但它们对于未见过的新生成模型泛化能力有限。这一局限主要归因于这些检测器依赖于特定生成方式产生的特征，如风格先验和压缩模式。 为了克服这些限制，我们提出了GAMMA框架，这是一个新型的训练方法，旨在减少领域偏差并增强语义一致性。GAMMA引入了多样化的操作策略，包括基于修复（inpainting）的操作和保持语义不变的扰动，以确保经过操作的内容与真实内容具有一致性。我们采用了多任务监督机制，并设置了两个分割头和一个分类头，这使得我们的模型能够在不同生成领域内实现像素级来源归属。此外，引入了一个反向交叉注意机制（reverse cross-attention mechanism），允许分割头指导并纠正分类分支中的偏差表示。 在GenImage基准测试中，我们的方法实现了最先进的泛化性能，准确率提高了5.8%，同时对于新发布的生成模型如GPT-4o也保持了强大的鲁棒性。
URL
https://arxiv.org/abs/2509.10250
