PGC: Peak-Guided Calibration for Generalizable AI-Generated Image Detection
For researchers and practitioners in AI-generated image detection, PGC provides a novel method to improve detection reliability by focusing on local discriminative clues, achieving significant gains over existing global approaches.
The paper tackles the problem of detecting AI-generated images, where subtle discriminative clues are often overshadowed by dominant image content. The proposed Peak-Guided Calibration (PGC) framework achieves state-of-the-art performance, improving mean accuracy by +12.3% on the CommGen15 dataset and setting new records on standard benchmarks.
The rapid evolution of generative AI, from GANs to modern diffusion models, has resulted in increasingly subtle discriminative clues. These fine-grained signals are often overshadowed by dominant, high-fidelity image content (e.g., the main subject), limiting the reliability of existing detectors that predominantly rely on global representations. To address this challenge, we propose the Peak-Guided Calibration (PGC) framework. PGC introduces a novel strategy that aggregates salient features via a peak-focusing mechanism. Specifically, by employing a peak-sensitive aggregation that accentuates the most discriminative local clues, PGC leverages these critical signals to calibrate the global decision. This approach recovers subtle patterns that would otherwise be submerged in the global context. Furthermore, to better simulate real-world threats, we introduce the CommGen15 dataset, a challenging benchmark comprising samples from 15 commercial models. Extensive experiments demonstrate that PGC achieves state-of-the-art performance. Specifically, it improves mean accuracy by +12.3% on our CommGen15 dataset, and sets new records on standard benchmarks, including GenImage (+2.1%), AIGI (+3.5%), and UniversalFakeDetect (+1.7%). Code is available at https://github.com/xiaoyu6868/PGC.