发表机构
College of Computer Science and Artificial Intelligence, Fudan University; Department of Information Engineering and Mathematics, University of Siena（复旦大学计算机科学与人工智能学院; 锡耶纳大学信息工程与数学系）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究无训练合成图像归因中表示与参考选择的相互作用，利用CLIP和DINOv2不同层表示及三种参考选择方法，发现归因准确性在中间表示层达峰值，语义约束参考可提升归因，重新合成在低参考量时有用，语义对齐参考在中等参考池时权衡较好。
AI 中文摘要
合成图像归因旨在识别给定人工智能生成图像的生成器。无训练的基于参考的归因方法易于扩展，因为新出现的生成器可通过添加特定源参考而非重新训练特定任务分类器来纳入。其性能取决于两个相互关联的因素：用于比较的表示空间和特定源参考的构建方式。本文利用参考和现成的预训练表示对这种相互作用进行了可控分析。研究了从CLIP和DINOv2不同层提取的表示，以及三种具有不同语义约束的参考选择方法：任意、语义对齐和基于重新合成的参考。结果表明归因准确性在中间表示层始终达到峰值，表明在强语义抽象主导之前，源区分线索更易获取。还表明中间表示并非完全语义中立，使参考选择至关重要：语义约束参考减少查询 - 参考不匹配并提高归因，尤其是在参考预算有限时。重新合成在低参考量情况下最有用，而语义对齐参考在有中等规模参考池时提供更好的准确性 - 成本权衡。研究结果表明无训练的基于参考的归因应理解为图像比较位置、参考集构建方式和可用参考数量之间的相互作用。
英文摘要
Synthetic image attribution aims at identifying the generator responsible for a given AI-generated image. Training-free reference-based attribution methods are easily scalable, since newly emerging generators can be incorporated by adding source-specific references rather than retraining a task-specific classifier. Their performance depends on two coupled factors: the representation space used for comparison and the way source-specific references are constructed. However, the interaction between these two factors remains largely unexplored. In this paper, we provide a controlled analysis of this interaction using references and off-the-shelf pretrained representations. We study representations extracted from different layers of CLIP and DINOv2, along with three reference selection methods with varying semantic constraints: arbitrary, semantically aligned, and resynthesis-based references. Our results show that attribution accuracy consistently peaks at intermediate representation levels, indicating that source-discriminative cues are more accessible before strong semantic abstraction dominates. We further show that intermediate representations are not completely semantically neutral, making reference selection critical: semantically constrained references reduce query-reference mismatch and improve attribution, especially under limited reference budgets. Resynthesis is most useful in low-reference regimes, while semantically aligned references provide a better accuracy-cost trade-off when a moderate-sized reference pool is available. Our findings show that training-free reference-based attribution should be understood as the interaction between where images are compared, how the reference set is constructed, and how many references are available.
Comments6 pages, 5 figures, 4 tables
