arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.23115cs.CV
收藏

AgentFoX: 基于可解释性的大语言模型引导融合的AI生成图像检测

AgentFoX: LLM Agent-Guided Fusion with eXplainability for AI-Generated Image Detection

Guangdong Provincial Key Laboratory of Intelligent Information Processing(广东省智能信息处理重点实验室);Shenzhen Key Laboratory of Media Security(深圳媒体安全重点实验室);SZU-AFS Joint Innovation Center for AI Technology(深圳大学-AFS人工智能技术联合创新中心);Shenzhen University(深圳大学);School of Cyber Science and Technology(网络安全科学与技术学院);The Smart City Research Institute of China Electronics Technology Group Corporation(中国电子科技集团有限公司智慧城市研究院)

Yangxin Yu, Yue Zhou, Bin Li, Kaiqing Lin, Haodong Li, Jiangqun Ni, Bo Cao

AI总结:

AgentFoX通过多阶段分析流程提升AI生成图像检测的可靠性,结合专家和上下文聚类资料,生成可读性强的报告以增强可解释性与可信度。

摘要:

随着AI生成图像的逼真度提升,亟需能可靠区分合成内容与真实图像的鉴伪工具。现有检测器通常针对特定伪造特征(如频域模式或语义不一致)进行优化,导致性能专业化且判断冲突。为解决这些问题,我们提出了AgentFoX框架,通过大语言模型引导的动态多阶段分析流程重新定义AIGI检测。该方法采用快速整合融合机制,结合校准的专家资料和上下文聚类资料。推理过程中,代理首先进行高层语义评估,再进入细粒度、上下文感知的信号层面专家证据合成,通过结构化推理解决矛盾。不同于返回粗略二元输出,AgentFoX生成详细的人可读鉴伪报告以支持其判决,提升可解释性和可信度。除提供新检测方案外,该工作还引入了可扩展的代理范式,促进未来和演进的鉴伪工具智能整合。

英文摘要:

The realism of AI-generated images (AIGI) poses increasing challenges for reliable forensic detection, where heterogeneous expert detectors may produce conflicting predictions across diverse generative sources and post-processing conditions. Existing multi-expert fusion methods rely on fixed rules or learned fusion strategies, offering limited ability to assess sample-specific reliability, execute rigorous adjudication of conflicts, and provide evidence-grounded explanations. We propose AgentFoX, an LLM-driven agentic multi-expert framework for AIGI detection that employs a command-and-reasoning core to perform evidence fusion. Following predefined guidelines, the core coordinates designated subtasks to collect semantic and signal-level evidence, reason over structured contexts to determine authenticity, and generate an auditable report for explainability. During this process, Expert Profiles are constructed for model-centric reliability assessment, while Clustering Profiles are built for data-centric contextual analysis, jointly establishing evidence contexts for conflict resolution. Extensive evaluations across diverse benchmarks demonstrate the robustness and generalizability of AgentFoX under complex conditions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑