发表机构
Baidu Inc.; Sun Yat-sen University; Tsinghua University（百度公司; 中山大学; 清华大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文构建文档图像检测基准AIGDoc-Pilot与AIGDoc，揭示现有AI生成图像检测器在文档场景下性能大幅下降（AUC降超7%），并发现空间不一致性与文本密度是关键因素，为开发可靠检测器提供见解。
AI 中文摘要
AI生成图像检测已引起越来越多的关注，但现有评估主要聚焦于自然图像，导致AI生成的文档图像在很大程度上未被充分探索。这一遗漏令人担忧，因为文档常出现在敏感的真实世界场景中，如发票、费用报告、证书和医疗记录。在本文中，我们首先构建了一个受控的诊断基准AIGDoc-Pilot，并揭示现有检测器在AI生成的文档图像上性能大幅下降，平均AUC下降超过7%。基于此，我们进一步揭示了这一差距背后的两个文档特有属性：生成伪影在局部区域间表现出强烈的空间不一致性，且文本密度显著影响真实与合成图像的可分性，其中文本密集区域提供更强的判别性证据。受这些发现的启发，我们构建了AIGDoc，一个更大的以文档为中心的数据集，包含多样化的真实世界文档及由多种先进生成和编辑模型产生的AI生成对应物。在AIGDoc上进行的大量实验表明，现有检测器仍难以可靠地识别AI生成的文档，而基于文档的训练部分缩小了这一差距。总之，这些结果为在文档中心场景中开发可靠且可泛化的检测器提供了宝贵的见解。代码和数据集将在论文被接收后公开提供。
英文摘要
AI-generated image detection has attracted increasing attention, but existing evaluations mainly focus on natural images, leaving AI-generated document images largely underexplored. This omission is concerning because documents often appear in sensitive real-world scenarios, such as invoices, expense reports, certificates, and medical records. In this paper, we first construct a controlled diagnostic benchmark, AIGDoc-Pilot, and reveal that existing detectors suffer substantial performance degradation on AI-generated document images, with the mean AUC dropping by more than 7%. Based on this, we further reveal two document-specific properties behind this gap: generation artifacts exhibit strong spatial inconsistency across local regions, and text density significantly affects real-synthetic separability, where text-dense regions offer stronger discriminative evidence. Motivated by these findings, we construct AIGDoc, a larger document-centric dataset containing diverse real-world documents and AI-generated counterparts produced by multiple advanced generation and editing models. Extensive experiments on AIGDoc demonstrate that existing detectors still struggle to reliably identify AI-generated documents, while document-based training partially narrows the gap. Together, these results offer valuable insights for developing dependable and generalizable detectors in document-centric scenarios. The code and datasets will be made publicly available upon acceptance of the paper.
