We are excited to highlight the groundbreaking work led by faculty member Assoc. Prof. Aykut Erdem, featuring the collaborative efforts of researchers at Koç University, Hacettepe University, and Adobe Research, recently accepted in ACM Transactions on Graphics. Their paper, titled ‘CLIP-guided StyleGAN Inversion for Text-driven Real Image Editing,’ explores the cutting-edge realm of Generative AI. This innovative research introduces a novel approach to real image editing, using natural language descriptions as a guide. Unlike previous methods, which had limitations in efficiency and handling multi-attribute changes, the team’s CLIPInverter stands out for its ability to efficiently and reliably perform complex image edits. Their approach achieves exceptional accuracy and photorealism across various domains, including human faces, cats, and birds. Dr. Erdem’s work has the potential to revolutionize image editing, offering new possibilities for creative applications across diverse fields.

Click the link below to access the article: