Perception, Verdict, and Evolution: Hindsight-Driven Self-Refining Forensics Agent for AI-Generated Image Detection

2026-06-24

AI Summary

Existing methods for detecting AI-generated images exhibit insufficient sensitivity to fine-grained forensic traces and rely heavily on static supervision, resulting in limited flexibility and high annotation costs. To address these limitations, this work proposes ForeAgent, a novel perception-adjudication framework that integrates multi-perspective cues from semantic, spatial, and frequency domains. The approach introduces a hindsight-driven self-refining mechanism, which employs a sample-reflect-evolve paradigm to autonomously generate high-quality reasoning trajectories and iteratively improve detection performance. By synergistically combining multimodal large language models, multi-domain feature fusion, dual-expert quality gating, and self-supervised fine-tuning, ForeAgent achieves 82.18% accuracy on the Chameleon benchmark—surpassing AIDE by 16.41%—and attains an average accuracy of 93.3% across 16 generators on the AIGCDetect-Benchmark, substantially advancing both detection efficacy and causal consistency.

Abstract

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.

Problem

• AI-generated image detection
• deepfake detection
• forensic artifacts
• multimodal large language models
• synthetic supervision

Innovation

• Self-Refining Agent
• Perception-Verdict Architecture
• Hindsight-Driven Learning
• AI-Generated Image Detection
• Multimodal LLM