DocumentCode
2554629
Title
Motion registration and correction based iterative reconstruction method for instant CT
Author
Ming Chang ; Liang Li ; Zhiqiang Chen ; Yongshun Xiao ; Ge Wang
Author_Institution
Dept. of Eng. Phys., Tsinghua Univ., Beijing, China
fYear
2012
fDate
Oct. 27 2012-Nov. 3 2012
Firstpage
2366
Lastpage
2369
Abstract
Dynamic CT aiming at reconstructing an object with time-dependent transformation is an important direction for the future development both in medical and industrial fields. Instant CT as a promising mode for dynamic CT is adopted in this work. It makes it possible to obtain the projection data of several views simultaneously for each motion phase using the multisource and multidetector configuration, though the X-ray source scanning remains indispensable for the complete data collection. However obvious artifacts are often introduced into the reconstructed image using TV minimization based methods due to the inadequate amount of projections in each phase. In this paper, a motion correction based iterative reconstruction method is purposed to overcome this problem. The main advantage of our method is taking full use of all the projections to achieve a high resolution reconstructed image which can be used as a prior image in the following PICCS process. The local deformable objects can also be well recovered in each motion phase applying PICCS method. Specifically our method can be divided into the following steps: firstly, a set of low spatial resolution images are reconstructed using the limited projection data, and then the global rigid motion can be estimated from these images incorporating the SIFT method; secondly, a high spatial resolution image is achieved by a motion correction based iterative algorithm using all the projection data; finally, the local blurs caused by the non-rigid motion at different motion phase are eliminated from the dynamic image sequences by PICCS method taking the high spatial resolution image as a prior image. The numerical experiment results show that the spatial resolution is significantly improved while the local deformable objects are well reconstructed.
Keywords
computerised tomography; deformation; image reconstruction; image registration; image sequences; iterative methods; medical image processing; motion compensation; PICCS method; SIFT method; TV minimization based method; X-ray source scanning; data collection; data projection; dynamic computed tomography; dynamic image sequence; industrial field; iterative reconstruction method; local deformable object reconstruction; medical field; motion correction; motion phase; motion registration; multidetector configuration; multisource configuration; numerical experiment; spatial resolution image; time-dependent transformation;
fLanguage
English
Publisher
ieee
Conference_Titel
Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), 2012 IEEE
Conference_Location
Anaheim, CA
ISSN
1082-3654
Print_ISBN
978-1-4673-2028-3
Type
conf
DOI
10.1109/NSSMIC.2012.6551538
Filename
6551538
Link To Document