臺灣博碩士論文加值系統

醫療影像可分為成電腦斷層掃描簡稱為CT、磁振造影簡稱為MRI、單光子斷
層掃描簡稱為SPECT、正子斷層掃描簡稱為PET。PET 為目前醫學上診斷癌症、
神經精神疾病及心臟病上具有相當重要性的診斷工具。PET 早期的影像重建方式
多以濾波反投影法簡稱FBP，目前大多以簡稱OSEM 之序列子集迭代式影像重建
法，來進行加速重建影像為最佳的方式。平行處理是為了解決單一處理器計算能
力不足而被提出來的。如果使用平行處理來加速OSEM 演算法，將可以大量減少
OSEM 運算之時間。平行處理必須使用多處理器系統，在實作時需克服許多問題，
例如：工作的分配、工作的切割、資料獨立性考量等，所以本研究利用迭代運算
與迭代運算間的特性以及一個迭代運算內的特性，提出Intra-iteration OSEM、
Inter-iteration OSEM、Intra & Inter-iteration OSEM 三種利用平行運算來
加速OSEM。
OpenMP 也是在共享記憶體多處理器系統上最常用的程式介面，為考慮資料
的特性，因此本研究選用OpenMP 來完成平行程式的實作。這些演算法可以與原
始OpenMP 加速矩陣運算的方式一起使用，若將OpenMP 的設計與此研究設計的
OSEM 加速方式混用，將可以大量的減少OSEM 運算的時間。

Medical imaging techniques include computerized tomography (CT),
magnetic resonance imaging (MRI), single photo emission computerized
tomography (SPECT), and positron emission tomography (PET). PET is a
rather important tool in today’s medicine to diagnose cancer,
neuropsychological disease, and heart disease. In the early days, filtered
back-projection was mostly used to reconstruct PET images. Since then
OSEM has become the iterative algorithm most frequently used by SPECT
and PET imaging. A parallel process was developed to make up the deficiency
of merely using a single processor. Using parallel process is expected to
expedite OSEM algorithm will considerably reduce OSEM calculation time.
However, parallel process requires using multi-processing system, and, in
practice, it needs to overcome many problems, such as task distribution, task
segmentation, and data dependence. Thus, this study proposes the ways to
use parallel computation to speed up OSEM process. Based on the nature of
iterative algorithm and the characteristics of shared memory multiprocessor
system, the study proposes three methods. They are intra-iteration OSEM,
inter-iteration OSEM, and intra- & inter-iteration OSEM schemes all of which
are trying to use parallel calculation to accelerate OSEM.
iv
In practice, OpenMP has been used a lot to provide handy multi-execution
parallel API to work with the hardware of shared memory structure. Base on
the characteristics of data, the study will use OpenMP to implement the
parallelization.

誌謝................................................................i
中文摘要........................................................... ii
英文摘要........................................................... iii
目錄................................................................v
圖目錄............................................................. vi
1 研究動機與目的....................................................1
2 文獻探討..........................................................2
3 研究方法..........................................................6
3.1 Intra-iteration OSEM 改良方式...............................8
3.2 Inter-iteration OSEM 改良方式...............................9
3.3 Intra&Inter-iteration OSEM 改良方式........................10
3.4 實作設計...................................................12
4 實驗與結果.......................................................13
5 結論與建議.......................................................20
參考文獻...........................................................21

[1]John M. Ollinger and Jeffrey A. Fessler. “ Positron-emission tomography”. IEEE Signal Processing Magazine, Vol. 41, No.1, pp. 43-55, January 1997.
[2]S. Vandenberghe, Y. D’ Asseler, R. Van de Walle, T. Kauppinen, M. Koole, L. Bouwens, K. Van Laere, I. Lemahieu, and R. A. Dierckx. “Iterative reconstruction algorithms in nuclear medicine”. Computerized Medical Imaging and Graphics, Vol. 25, pp. 105-111 ,2001.
[3]L. A. Shepp and Y. Vardi. “ Maximum likelihood reconstruction in Positron Emission Tomography”. IEEE Trans. Med. Imag., vol. 1(2), pp. 113-122 ,1982.
[4]H.M. Hudson, R.S. Larkin. “ Accelerated image reconstruction using ordered subsets of projection data”. IEEE Trans. Med. Imag. 13,601, 1994.
[5]Cliff X. Wang, Wesley E. Snyder, Griff Bilbro, and Pete Santago. “ Performance evaluation of filtered backprojection reconstructionand iterative reconstruction methods for PET images”. Computers inBiology and Medicine, vol. 28, pp. 13-25 , 1998.
[6]M. Defrise, P. E. Kinahan, D.W. Townsend, C. Michel,M. Sibomana, and D. F. Newport. “ Exact and Approximate Rebinning Algorithms for 3-d PET Data”. IEEE Transaction on Medical Imaging, 16, April 1997.
[7]P.E. Kinahan and J.G. Rogers. (1990). Analytic three-dimensional image reconstruction using all detected events. IEEE Trans. Nucl. Sci. NS-36 964-8.
[8]Michael J.Quinn. (2003).Parallel Programming in C with MPI and OpenMP.New York:McGraw-Hill
[9]R. Chandra, L. Dagum, D. Kohr, D. Maydan, J. McDonald, R. Menon (2001). Parallel Programming in OpenMP, Morgan Kaufmann Publishers Inc.
[10]J.G. Colsher. (1980). Fully three-dimensional positron emission tomography. Physics in Medicine and Biology, 25(1),p. 103-115.
[11]Z.H. Cho, J.B. Ra, and S.K. Hilal. “True three-dimensional reconstruction (TTR) – Application of algorithm toward full utilization of oblique rays”. IEEE Trans. Med. Imag. MI-1(6), 6 , 1983.
[12]J.B. Ra, C.B. Lim, Z.H. Cho, S.K. Hilal, and J. Corell. . “ A true three-dimensional reconstruction algorithm for the spherical positron emission tomography”. Phys. In Med. And Biol. 27(1),37 ,1982.
[13]M. Defrise, S. Kuijk, and F. Deconinick. “ A new three-dimensional reconstruction method for positron cameras using plane detectors”. Phys. In Med. And Boil.33(1),43 ,1988.
[14]R. L. Siddon. “ Fast calculation of the exact radiological path for a three-dimensional CT array”. Medical Physics,12-2, p. 252-255 ,1985.
[15]Linda Kaufman. “Implementing and Accelerating the EM Algorithm for Positron Emission Tomography”. IEEE Trans. Med. Imag., vol.MI-6, no.1,pp.37-51 , 1987.
[16]C.A. Johnson, A. Sofer. “ A Data-Parallel Algorithm for Iterative Tomographic Image Reconstruction”. The 7th Symposium on the Frontiers of Massively Parallel Computation,pp.126 ,1999 .
[17]M. Bertero, P. Bonetto, L. Carracciuolo, L. D'Amore, A. Formiconi, M. R. Guarracino, G. Laccetti, A. Murli, G. Oliva. “ MedIGrid: a Medical Imaging application for computational Grids”. International Parallel and Distributed Processing Symposium, pp. 22-26 , April 2003 .
[18]Tao He, Jun Ni, Ge Wang. A Heterogeneous Windows Cluster System for Medical Image Reconstruction”. First International Multi-Symposiums on Computer and Computational Sciences, pp.410-415 , June 2006.
[19]Laksono Adhianto, Barbara Chapman. “Performance Modeling of Communication and Computation in Hybrid MPI and OpenMP Applications”. 12th International Conference on Parallel and Distributed Systems - Volume 2, pp. 3-8 , July 2006 .
[20]Xuan Liu,Claude Comtat,Christian Michel,Paul Kinahan,Michel Defrise*, and David Townsend.“Comparison of 3-D Reconstruction With 3D-OSEM and With FORE+OSEM for PET”. IEEE Trans. Med. Imag., vol.20 no.8. pp. 804-814 , August 2001 .
[21]林暉涵。平行處理在醫學影像處理上的應用。國立清華大學原子科學研究所碩士論文，2001年。