AgentFoX: LLM Agent-Guided Fusion with eXplainability for AI-Generated Image Detection
About
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.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| AI-generated image detection | WildFake (All) | F1 Score93.26 | 32 | |
| AI-generated image detection | X-Fuse | Overall (F1)80.87 | 21 | |
| AI-generated image detection | WIRA All | F1 Score95.79 | 17 | |
| AI-generated image detection | WIRA and WildFake Overall | F1 Score94.88 | 17 | |
| Cross-domain AI-generated image detection | AIGCDetect | F1 Score75.58 | 15 | |
| Cross-domain AI-generated image detection | AIGIBench | F1 Score89.75 | 15 | |
| Cross-domain AI-generated image detection | Chameleon | F1 Score69.14 | 15 | |
| Cross-domain AI-generated image detection | Community Forensics | F1 Score91.5 | 15 | |
| Cross-domain AI-generated image detection | WIRA | F1 Score73.23 | 15 | |
| Cross-domain AI-generated image detection | GenImage | F1 Score78.54 | 15 |