조천식녹색교통대학원 공승현교수연구실에서
11월 22일(목) AM 10:30~12:00,
Dr. Bongyong Song (Sr. Staff Engineer at Qualcomm Research Center, San Diego)의 "Compressed Sensing을 이용한 차세대 CT 기술" 세미나를 KI Bldg. Faculty Conference Room (3F)에서 아래와 같이 개최하오니, 관심있는 분들의 많은 참석 바랍니다.
= 아 래 =
– Title : Next Generation Computed Tomography Using Compressed Sensing – Fast, Iterative, Low-dose, Cone-Beam Computed Tomography (CBCT) Reconstruction
– Speaker : Dr. Bongyong Song (Sr. Staff Engineer at Qualcomm Research Center, San Diego)
– Abstract:
In order to minimize the potential health risks due to ionizing radiation scans, it is important to use the lowest optimal radiation dose in computed tomography procedures. Recent advancements in the compressive sensing theory provide machinery for enabling CT/CBCT reconstructions with fewer and/or noisier projection data. In this talk, I will first review the computed tomography principles and explain the inherent sparsity in the images. This sparsity allows us to formulate the CT/CBCT reconstruction problem as a Total Variation (TV) regularized least squares problem where the sparsity is promoted by the TV regularization term. To solve this problem, we propose a novel gradient descent method based on the Barzilai and Borwein (BB) approach to enable fast, low-dose, iterative CBCT reconstructions. By implementing the proposed algorithm on an off-the-shelf GPU hardware, we demonstrate that the proposed algorithm reconstructs the CBCT images in an extremely efficient manner, with speed and quality acceptable for use in on-line image-guided radiation therapy (IGRT). Extensions to 4D-CBCT will also be discussed.
– Biography :
Bongyong Song is a Sr. Staff Engineer at Qualcomm Research Center, San Diego. He received the B.S. and M.S. degrees in Electronic Engineering from Sogang University, Seoul, Korea, in 1994 and 1996, respectively, and the Ph.D. degree in Electrical and Computer Engineering from the University of California at San Diego, La Jolla, CA, in 2006. From 1996 to 2001, he was with the Research and Development Center, SK Telecom Co. Ltd., Korea where he worked on CDMA radio networks. Since 2006, he has been with Qualcomm Technology Inc., San Diego, CA, and he designed a commercially successful push-to-talk system over 3G/4G wireless systems, QChat. He is now leading the project that develops intelligent wireless devices and networks. His research interests include design, performance analysis and optimization of wireless networks, adaptive systems, and signal processing for communications. He is also interested in signal processing for medical physics with emphasis on medical imaging and applications. He is currently serving as an Assistant Professor (voluntary) in School of Medicine at University of California, San Diego. He is a recipient of the 2008 Stephen O. Rice paper prize award for the best original journal article in the IEEE Transactions on Communications.
– Time : 2012. 11. 22(Thu), AM 10:30 ~ 12:00
– Host : Prof. Seung-Hyun Kong
– Place : KI Bldg. Faculty Conference Room (3F)
– Contact Us : 안명순 (T. 042-350-1295/freezia7@kaist.ac.kr)
