报告题目：Improving Generalization in Image Restoration via ‘Unconventional' Losses

报告人：Chen Change Loy 教授 新加坡南洋理工大学

报告时间：2024年6月27日（周四），上午10:00—12:00

腾讯会议：823 222 870

报告摘要：In this talk, we explore strategies to enhance the generalization of image restoration and enhancement methods, focusing on the use of 'unconventional' losses. Traditional deep learning-based image restoration techniques often struggle with real-world scenarios due to domain gaps between synthetic and real-world data. We will discuss two new approaches that address this challenge. The first approach uses Contrastive Language-Image Pre-Training (CLIP) for unsupervised backlit image enhancement, employing a prompt learning framework to optimize the enhancement network. The second approach leverages a novel domain adaptation strategy in the noise-space using diffusion models, incorporating contrastive learning to align restored images to a clean distribution. By integrating these "unconventional" loss functions, we demonstrate significant improvements in the visual quality and generalization ability of image restoration and enhancement tasks.

个人简介：Chen Change Loy is a President's Chair Professor with the College of Computing and Data Science, Nanyang Technological University, Singapore. He is the Lab Director of MMLab@NTU and Co-associate Director of S-Lab. He received his Ph.D. (2010) in Computer Science from the Queen Mary University of London. Prior to joining NTU, he served as a Research Assistant Professor at the MMLab of The Chinese University of Hong Kong, from 2013 to 2018. His research interests include computer vision and deep learning with a focus on image/video restoration and enhancement, generative tasks, and representation learning. He serves as an Associate Editor of the International Journal of Computer Vision (IJCV), IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) and Computer Vision and Image Understanding (CVIU). He also serves/served as an Area Chair of top conferences such as ICCV, CVPR, ECCV, ICLR, and NeurIPS. He will serve as the Program Co-Chair of CVPR 2026. He is a senior member of IEEE.