Perception, Verdict, and Evolution: Hindsight-Driven Self-Refining Forensics Agent for AI-Generated Image Detection
Quick Answer
ForeAgent is a novel forensics framework for AI-generated image detection, achieving 82.18% accuracy on the Chameleon benchmark, outperforming AIDE by 16.41%.
Quick Take
ForeAgent is a novel forensics framework for AI-generated image detection, achieving 82.18% accuracy on the Chameleon benchmark, outperforming AIDE by 16.41%. It employs a Perception-Verdict architecture and a Hindsight-Driven Self-Refining strategy for continual self-improvement, demonstrating superior reasoning consistency compared to GPT-5.
Key Points
- ForeAgent uses a Perception-Verdict architecture to integrate multi-view cues.
- The framework achieves 93.3% mean accuracy on the AIGCDetect-Benchmark across 16 generators.
- It continuously evolves through a dual-expert quality gating module for high-quality sample curation.
- Extensive experiments confirm ForeAgent's state-of-the-art performance in deepfake detection.
- The model reflects on failure cases to enhance reasoning quality over time.
Paper Resources
📖 Reader Mode
~2 min readAbstract:The rapid advancement of generative models presents a significant challenge to existing deepfake detection methods, particularly given the widespread dissemination of highly realistic AI-generated images. Although Multimodal Large Language Models (MLLMs) show strong potential for this task, existing approaches suffer from two key limitations: insufficient sensitivity to fine-grained forensic artifacts and reliance on static synthetic supervision from frontier models, leading to limited flexibility and high-cost. To address these issues, we propose ForeAgent, an agentic forensics framework for AI-generated image detection with iterative self-evolution. First, ForeAgent adopts a Perception-Verdict architecture that aggregates multi-view cues spanning semantic, spatial, and frequency-domain features, and leverages an MLLM as a verdict module to fuse these signals for a logical-grounded verdict. Second, to enable continual self-improvement, we introduce a Hindsight-Driven Self-Refining strategy following a Sampling-Reflection-Evolution paradigm. The agent performs inference rollouts on training instances. Guided by ground-truth labels as hindsight, it reflects on failure cases and low-quality reasoning trajectories to regenerate higher-quality reasoning traces. These synthesized samples are then strictly filtered through a dual-expert quality gating module. ForeAgent continuously evolves via fine-tuning on self-curated high-quality samples. Extensive experiments demonstrate that ForeAgent achieves state-of-the-art performance on the Chameleon benchmark, reaching 82.18% accuracy (+16.41% over AIDE), and achieves 93.3% mean accuracy on AIGCDetect-Benchmark across 16 generators. In addition, external evaluation shows that ForeAgent produces more consistent and causally grounded reasoning compared to GPT-5 and GPT-5-mini.
| Comments: | 10 pages |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2606.26552 [cs.CV] |
| (or arXiv:2606.26552v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2606.26552 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yangjun Wu [view email]
[v1]
Thu, 25 Jun 2026 02:59:33 UTC (534 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →LLM-Guided ANN Index Optimization for Human-Object Interaction Retrieval
A phase-aware LLM agent optimizes human-object interaction retrieval, outperforming Optuna TPE by 33.3% and VDTuner by 34.2% on the HICO-DET benchmark. This method enhances throughput by 15.3x over UniIR and demonstrates strong transferability across vector database management systems.