Generative High-Magnification Image Synthesis for EUV Mask Inspection

This paper proposes a mask inspection image enhancement method based on the pix2pix architecture to synthesize high-magnification images from low-magnification images of EUV semiconductor pattern masks. To the best of our knowledge, this is the first work leveraging generative models to transform low-magnification EUV mask pattern images into their high-magnification equivalents. The proposed method improves defect detection accuracy while reducing the inspection costs associated with high-resolution imaging.