Order within Chaos: Capturing Intrinsic Energy Anomalies for AI-Manipulated Image Forgery Localization
For digital forensics, this work provides a theoretically grounded method to detect and localize forgeries from modern generative models, which are increasingly realistic and evade traditional techniques.
The paper addresses the challenge of localizing AI-manipulated image forgeries by leveraging the statistical energy gap between diffusion-generated and natural images. The proposed FLAME framework achieves state-of-the-art performance on AI-generated forgery datasets, significantly outperforming prior methods.
Recent advancements in generative AI have led to image editing models capable of producing realistic forgeries that evade traditional image forgery localization methods, as these approaches depend on physical noise absent in synthetic data. To address this challenge, we theoretically demonstrate that the diffusion process inherently suppresses local high-frequency variance, creating a statistical energy gap that is distinguishable from the natural entropy of optical imaging. Guided by this insight, we propose FLAME, a unified framework that utilizes a LAD map to capture these intrinsic anomalies, coupled with a parameter-efficient adapter for SAM to achieve precise, pixel-level forgery localization. Furthermore, to bridge the lag between forensic benchmarks and evolving generative models, we introduce EditStream, an automated pipeline for continuous, instruction-based training data synthesis. Extensive experiments demonstrate that FLAME establishes a new state-of-the-art, significantly outperforming previous methods on AI-generated forgery datasets while effectively generalizing to unseen generative architectures. Our code is available at https://github.com/phoenixnir/FLAME.