ForensicConcept: Transferable Forensic Concepts for AIGI Detection

2026-06-05

AI Summary

Current AI-generated image detectors suffer from poor generalization on out-of-distribution data and lack interpretability. This work proposes the first transferable and interpretable forensic concept framework: it localizes critical image patches via Transformer attribution, constructs a concept codebook through clustering, and achieves cross-architecture concept alignment and transfer by integrating generation artifact features extracted by CleanDIFT with neighborhood structural consistency measured by CKNNA. The method consistently outperforms existing approaches on GenImage, GAN-family, and Chameleon benchmarks. Notably, the CKNNA alignment score effectively predicts transfer performance, providing auditable evidence to support detection decisions.

Abstract

AI-generated image detectors achieve high accuracy on in-distribution data but often fail on unseen generators. A key obstacle to understanding this failure is the black-box nature of current detectors: they do not reveal which evidence drives their decisions. We propose ForensicConcept, a framework that extracts explicit forensic concepts from detectors and enables their transfer across backbones. Our method localizes decision-critical patches via Transformer attribution, clusters them into a compact concept codebook, and uses a concept-aligned projection to produce auditable evidence readouts. Motivated by prior studies showing that DINO representations can guide diffusion generation and exhibit concept-level correspondence with diffusion features, we introduce a generation-trace reference based on CleanDIFT diffusion features and quantify backbone-trace alignment via neighborhood-structure consistency (CKNNA). We further propose concept codebook injection to transfer diffusion-derived concepts into target backbones. Experiments on GenImage, GAN-family, and Chameleon benchmarks show consistent improvements over prior methods. We also find that CKNNA alignment predicts transfer effectiveness, providing a principled explanation for why some backbones yield more transferable forensic evidence than others.

Problem

• AIGI detection
• forensic concepts
• transferability
• black-box detectors
• unseen generators

Innovation

• ForensicConcept
• concept transfer
• diffusion features
• CKNNA alignment
• auditable detection