Recent advances in deep generative models have led to the development of
methods capable of synthesizing high-quality, realistic images. These models
pose threats to society due to their potential misuse. Prior research attempted
to mitigate these threats by detecting generated images,