Order within Chaos: Capturing Intrinsic Energy Anomalies for AI-Manipulated Image Forgery Localization
2026-06-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors explain that new AI tools can create fake images that are very hard to detect with older methods because those methods look for physical noise not present in AI-made images. They show that the way the AI generates images reduces certain small details differently than natural photos, creating a unique pattern that can be spotted. Using this idea, they create FLAME, a system that finds these hidden clues in images to locate forgeries very accurately. They also build EditStream, a tool to keep training FLAME with up-to-date fake images as AI improves. Their tests show FLAME works better than earlier techniques and can adapt to new AI image generators.
generative AIimage forgery localizationdiffusion processhigh-frequency variancestatistical energy gapLAD mapparameter-efficient adapterSAM (Segment Anything Model)EditStreaminstruction-based training data synthesis
Authors
Yiming Wang, Baiqi Wu, Qingming Li, Jiahao Chen, Tong Zhang, Shouling Ji
Abstract
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.