Image restoration or deconvolution that is to recover the original clean image from a noisy or corrupted image has long been an important and fundamental problem in image processing and computer vision. In this talk, a very effective deep learning-based image restoration techniques will be addressed, which are based on very deep convolutional networks. Increasing the network depth enlarges the receptive field, resulting a significant improvement in accuracy. By cascading small filters many times in deep network structures, contextual information over large image regions is exploited in an efficient way. With very deep networks, however, convergence speed becomes a critical issue during training. Simple yet effective training procedures for successful convergence of very deep networks will be discussed including residual learning and adjustable gradient clipping. We show that our proposed methods outperform existing deconvolution methods quantitatively as well as qualitatively.
Kyoung Mu Lee received the B.S. and M.S. Degrees in Control and Instrumentation Eng. from Seoul National University (SNU), Seoul, Korea in 1984 and 1986, respectively, and Ph. D. degree in Electrical Engineering from the University of Southern California in 1993. He is currently with the Dept. of ECE at Seoul National University as a professor. Prof. Lee has received several awards, in particular, the Most Influential Paper over the Decade Award by the IAPR Machine Vision Application in 2009, the ACCV Honorable Mention Award in 2007, the Okawa Foundation Research Grant Award in 2006, the Distinguished Professor Award from the college of Engineering of SNU in 2009, and both the Outstanding Research Award and the Shinyang Engineering Academy Award from the College of Engineering of SNU in 2010. He is currently serving as an AEIC (Associate Editor in Chief) of the IEEE TPAMI, an Area Editor of the CVIU, and has served as an Associate Editor of the IEEE TPAMI, the Machine Vision Application Journal and the IPSJ Transactions on Computer Vision and Applications, and the IEEE Signal Processing Letter. He also has served (or will serve) as a General Chair of ICCV2019, ACM MM2018, Program Chair of ACCV2012, a Track Chair of ICPR2012, Area Char of CVPR2012, CVPR2013, CVPR2015, ICCV2013, ECCV2014, ECCV2016 and a Workshop Chair of ICCV2013. He was a Distinguished Lecturer of the Asia-Pacific Signal and Information Processing Association (APSIPA) for 2012-2013. More information can be found on his homepage http://cv.snu.ac.kr/kmlee .