面向泛化AI生成图像检测的生成签名监督学习
Enabling Supervised Learning of Generative Signatures for Generalized AI-Generated Images Detection
用随机架构与参数的动态重建器在真实图像上模拟生成痕迹，以重建残差作为代理监督训练痕迹提取器，为缺乏真值的取证类任务（如伪造检测、篡改定位）提供了一条可复用的监督构造思路。
原文依据 The reconstruction residuals serve as supervision to train an extractor that learns to isolate traces
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论从生成图像中提取可靠的生成痕迹对AI生成图像（AIGIs）检测至关重要。然而，存在一个根本性挑战：AIGIs天然包含生成痕迹，却不存在无痕迹的对应样本，这使得对这些伪影进行监督式提取不可行。本工作通过代理监督框架克服这一难题。我们设计了一个动态重建器，通过随机变化的架构与参数在真实图像上模拟多样化的生成痕迹。重建残差作为监督信号，用于训练一个学习分离痕迹（即生成签名，GenSign）的提取器。随后检测器将提取的GenSign与RGB特征融合，以区分真实图像与AIGIs。我们的关键洞见是：模拟中充分的架构多样性能够实现对真实世界生成器的有效迁移，从而解决真值GenSign缺失的问题。在四个基准上的大量实验证明了最优的泛化能力，证实我们基于模拟的学习范式能够提取通用且可迁移的取证特征。
英文原文摘要 1148 字符
Extracting reliable generative traces in generated images is critical for AI-generated images (AIGIs) detection. However, a fundamental challenge exists: AIGIs inherently contain generative traces with no trace-free counterpart available, making supervised extraction of these artifacts infeasible. In this work, we overcome this through a surrogate supervision framework. We design a dynamic reconstructor that simulates diverse generative traces on real images through stochastically varied architectures and parameters. The reconstruction residuals serve as supervision to train an extractor that learns to isolate traces, i.e., generative signatures (GenSign). A detector then fuses extracted GenSign with RGB features to distinguish real images from AIGIs. Our key insight is that sufficient architectural diversity in simulation enables effective transfer to real-world generators, resolving the absence of ground truth GenSign. Extensive experiments across four benchmarks demonstrate state-of-the-art generalization, confirming that our simulation-based learning paradigm is capable of extracting general and transferable forensic features.