Jingyi Hou

Progressive Decision-Making for Localizing Open-Ended AI-Generated Image Forgeries

Jingyi Hou, Xiaoxia Chen, Leyu Zhou, Zhichuang Wang, Zhijie Liu

AI-generated image forgeries are becoming increasingly realistic and difficult to characterize with fixed manipulation patterns. As generative models continue to evolve, it is impractical to expect a localization model to exhaustively learn all possible forgery appearances from l…Read more