[3Win5] Poster session 3

[3Win5-46] Investigation of the Effectiveness of Focusing on Local Regions in the Detection of Diffusion Model Generation Image

〇Takumi Owada¹, Kenya Jin'no¹ (1.Tokyo City University)

Keywords:Diffusion Models, AI-Generated Images, CNN

In recent years, image generation techniques using diffusion models have rapidly advanced. Their flexibility, realism, and ease of implementation have raised ethical concerns, such as the spread of fake news. Conventional approaches for detecting these generated images typically rely on applying computationally expensive neural networks to the entire high-resolution image. However, our previous research revealed that partitioning the image and focusing on local regions can achieve both higher detection accuracy and lower computational load. In this study, we investigate why such detection methods are effective and conduct a detailed examination of the characteristics of images generated by diffusion models.