学 术 讲 座

题   目（TITLE）：Self-Learning-Based Structured Noise Removal in a Single Image

讲座 人（SPEAKER）：Chia-Wen Lin  (National Tsing Hua University)

主持 人   (CHAIR)： Jinqiao Wang

时  间    (TIME)： Jul 9, 2013(Tuesday), 14:00-16:00

地  点   (VENUE)：No.2 Conferece Room (3rd floor), Intelligence Building

报告摘要（ABSTRACT）：

Decomposition of an image into multiple semantic components has been an effective research topic for various image processing applications, such as image denoising, enhancement, and inpainting. In this talk, we introduce a method of image decomposition based on the uses of sparsity and morphological diversity in image mixtures. First, we analyze existing MCA (morphological component analysis) based image decomposition frameworks with their applications and explore the potential limitations of these approaches for image denoising. Then, we discuss our recently proposed self-learning based image decomposition framework with its applications to several image denoising tasks, including single image rain streak removal, denoising, super-resolution, and deblocking for a highly compressed image/video. By advancing sparse representation and morphological diversity of images, the proposed framework first learns an over-complete dictionary from the high spatial frequency parts of an input image for reconstruction purposes. An unsupervised clustering technique is applied to the dictionary atoms for identifying the morphological component corresponding to the noise pattern of interest (e.g., rain streaks, blocking artifacts, or Gaussian noise). The proposed self-learning based approach allows one to identify and disregard the above morphological component during image reconstruction in an unsupervised way, and thus image denoising can be achieved automatically.  Different from prior learning or MCA-based image decomposition works, the proposed method does not need to collect training data in advance. Our experimental results have confirmed the effectiveness and robustness of our framework, which is shown to outperform state-of-the-art approaches.

报告人简介（BIOGRAPHY）：

Chia-Wen Lin received his Ph.D. degree from the Department of Electrical Engineering, National Tsing Hua University (EE/NTHU), Hsinchu, Taiwan, in January 2000. He has been with EE/NTHU as Associate Professor since August 2007. Prior to joining EE/NTHU, he worked for the Department of Computer Science and Information Engineering, National Chung Cheng University (CSIE/CCU), Chiayi, Taiwan from August 2000 to July 2007. He was with the Information and Communications Research Laboratories, Industrial Technology Research Institute (ICL/ITRI), Hsinchu, Taiwan, during 1992-2000, where his final post was Section Manager. His research interests include video networking and video content analysis. Dr. Lin was awarded the Young Faculty Awards presented by National Chung Cheng University in 2005 and the Ta-You Wu Memorial Awards presented by National Science Council (NSC), Taiwan in 2006. His paper won the Young Investigator Award presented by SPIE VCIP 2005. He serves on the editorial board of IEEE Transactions on Circuits and Systems for Video Technology, IEEE Transactions on Multimedia, IEEE Multimedia Magazine, Journal of Visual Communication and Image Representation, and Signal Processing: Image Communication. He serves as Chair-Elect of Multimedia Systems & Applications of IEEE Circuits and Systems Society. He was TPC Co-Chair of IEEE International Conference on Multimedia & Expo (ICME) 2010 and Special Session Co-Chair of IEEE ICME 2009.