報告內容：Detail-preserving face photo-to-sketch synthesis
報告人：Prof. HYO JONG LEE
報告時間🏭：10月30日 15:00
報告地點🚶🏻：現代交通工程中心7950會議室
報告人簡介:
Professor Lee received B. S. (1986), M. S. (1988) and Ph. D. (1991) in Computer Science from University of Utah, Salt Lake City, Utah. He also received another B. S. (1982) from Jeonbuk National University and M. S. (1985) from University of Utah in Meteorology. Dr. Lee is a professor at the Division of Computer Science, Jeonbuk National University since 1991. He is also a president of company called “AI Tech”. He has been a director of the Center for Advanced Image and Information Technology from 2007~2019.
Professor Lee’s research interests include image processing, artificial intelligence, parallel algorithms and bioinformatics. He is a member of ACM and IEEE. Currently he is leading multiple projects including “Deep Network-based Face Sketch Recognition”, “Prediction of Ovulatory Period of Mother Pigs”, “Face Recognition under Hazard Condition”, and “Construction of Geo-spatial Database for Korean Village Community.”
During the past five years, Professor Lee published 43 prestigious journal papers and 100 conference papers.He also registered five Korean patents and one international patents.
報告內容簡介:
Face sketch synthesis aims to generate a face sketch image from a corresponding photo image and has wide applications in law enforcement and digital entertainment. Despite the remarkable achievements that have been made in face sketch synthesis, most existing works pay main attention to the facial content transfer, at the expense of facial detail information. In this paper, we present a new generative adversarial learning framework to focus on detail preservation for realistic face sketch synthesis. The main architecture of our method is based on generative adversarial network (GAN), in which a modified high-resolution network is adopted as generator to transform a face image from photograph to sketch domain. Except for the common adversarial loss, we design a detail loss to restrain the synthesized face sketch image to have proximate details to its corresponding photo image. Experimental results indicate that the proposed approach achieves superior performance, compared to state-of-the-art approaches, both on visual perception and objective evaluation.
