核磁共振造影(Magnetic Resonance Imaging, MRI)利用磁場原理，在均勻的靜磁場中，利用無線電波脈衝，使人體內水分子中的氫原子產生共振，進而改變磁場所產生的回波訊號，由電腦組合成影像。經過長期臨床實驗，已證實對人體無害，由於其非侵入性與無輻射傷害的優點，在同一切面下用不同頻率去掃描，會得到不同的強度特徵，配合強大的成像序列變化，使其提供了龐大且豐富的組織訊息，相對的卻造成醫師病理判讀的負擔，如何使用電腦計算來克服龐大資料量所帶來困擾，成為本文研究重點。
本文應用自動向量種子區域成長(Automatic Vector Seeded Region Growing, AVSRG)來對腦部MRI進行組織分類，透過自動挑選向量種子，依種子進行區域成長，將腦部進行細部分割，最後將分割後區域用K-means進行分類，並以單張影像分別呈現大腦灰質(Gray Matter)、白質(White Matter)及腦脊髓液(Cerebral Spinal Fluid)三大組織，在以往醫師判讀影像憑著醫學知識，現在透過自動向量種子區域成長，供年輕醫師為參考依據，對病患是雙重保障，並大幅減輕因MRI多頻譜龐大資訊對醫師帶來之負擔，使醫師診斷時更有效率，準確地判斷病灶之所在。為了證明AVSRG分類成效，以K-means和腦部自動分割工具(FMRIB's Automated Segmentation Tool, FAST)進行比較，並採用接收器操作特性(Receiver Operating Characteristic, ROC)分析來進行效能評估，並證明其優越性。

Making use of magnetic field theory, magnetic resonance imaging (MRI) which employs pulses radio waves and intense magnetic fields to vibrate hydrogen atoms in the body is a method of obtaining computer images. After long-term clinical trials, MRI has been proved to use in humans harmlessly because of its advantages of non-invasive and non-radiation. In addition, MRI provides cross-sectional images of the objective and scans with different frequency. In light of the procedure, it will generate the difference in strength characteristics, combining powerful imaging sequence changes to provide abundant organization information, which is the key to use computer algorithms to reduce the burden on physician interpretation. How to use computer to decrease processing time is the point, comparing to burden of the physicians diagnosis.
In this study, the automatic vector seed region growing is applied in the brain MRI classification. First, vector seed is selected automatically. Second, base on seed regional growth, brain is segmented in detailed. Finally, using K-mean to classify the brain segmentation images, according to gray matter, white matter, and cerebral spinal fluid. When physicians face many images, automatic analysis of medical images become the most challenging task. Nowadays automatic vector seed region growing which is proposed can be the reference for many researchers because it can decrease processing time particularly. This approach makes physicians’ diagnosis acute and effectiveness. Compare to K-means and FMRIB’s Automated Segmentation Tool, FAST, the classified statistics of AVSRG is above all else. To prove the automatic classification results Vector seeds of regions growing, using the receiver operating characteristic analysis to evaluate classification performance, and prove its superiority.

中文摘要 i
ABSTRACT ii
誌謝 iv
目錄 v
表目錄 vii
圖目錄 viii
第一章、緒論 1
1.1 研究動機 1
1.2 研究目的與方法 4
1.3 文獻回顧 6
1.4 章節架構 10
第二章、核磁共振造影(Magnetic Resonance Imaging) 11
2.1 簡介 11
2.2 MRI成像原理 13
2.3 影像資料來源 16
2.3.1 真實腦部MRI影像 18
2.3.2 假造橢圓MRI影像 21
2.3.3 BrainWeb-Simulated Brain Database 25
第三章、種子區域成長理論(Seeded Region Growing) 28
3.1 簡介 28
3.2 種子區域成長演算法 31
3.3 自動種子選擇法 34
第四章、自動向量種子區域成長 36
4.1 多頻譜影像處理 36
4.2 自動向量種子區域成長於MRI之分類 39
4.2.1 自動向量種子選擇 41
4.2.2 改良式區域成長 44
第五章、實驗結果與效能評估 46
5.1 假造橢圓MRI分類 47
5.2 BrainWeb : Simulated Brain Database分類 52
5.3 真實腦部MRI影像分類 57
5.4 接收器操作特性(Receiver Operating Characteristic, ROC) 58
第六章、結論 66
參考文獻 67

[1] 邱淑媞，「民國96年癌症登記報告」，行政院衛生署國民健康局。
[2] 認識磁振造影MRI，台中榮總磁振造影中心網站，網址：http://www.healthtc.com/。
[3] C. M. Wang, G. C. Lin, C. Y. Lin, and R. M. Chen, “An Unsupervised Kalman Filter-Based Linear Mixing Approach to MRI Classification,” Proceedings of the 2004 IEEE Asia-Pacific Conference on Circuits and Systems, vol. 2, pp. 1105-1108, 6-9 Dec. 2004.
[4] C. M. Wang, C. C. C. Chen, S. C. Yang, P. C. Chung, Y. N. Chung, C. W. Yang, and C. I. Chang, “An Unsupervised Orthogonal Subspace Projection Approach to MR Image Classification,” Optical Engineering, vol. 41, no. 7, pp. 1546-1557, Jul. 2002.
[5] C. M. Wang, S. C. Yang, P. C. Chung, C. I. Chang, C. S. Lo, C. C. Chen, C. W. Yang, and C. H. Wen, “Orthogonal Subspace Projection-Base Approaches to Classification of MR Images Sequences,” Computerized Medical Imaging and Graphics, vol. 25, no. 6, pp. 465-476, Dec. 2001.
[6] M. C. Clark, L. O. Hall, D. B. Goldgof, L. P. Clarke, R. P. Velthuizen, and M. S. Silbiger, “MRI Segmentation using Fuzzy Clustering Techniques,” IEEE Engineering Med. Biol. Mag., pp. 730-742, Dec. 1994.
[7] M. N. Ahmed, S. M. Yamany, A. A. Farag, and T. Moriarty, “Bias Field Estimation and Adaptive Segmentation of MRI Data using a Modified Fuzzy C-means Algorithm,” IEEE Computer Society Conference on Computer Vision and Pattern Recognition. vol. 1, pp. 250-255, Jun. 1999.
[8] M. N. Ahmed, S. M. Yamany, N. Mohamed, A. A. Farag, and T. Moriarty, “A Modified Fuzzy C-Means Algorithm for Bias Field Estimation and Segmentation of MRI Data,” IEEE Trans. on Medical Imaging, vol. 21, no. 3, pp. 193-199, Mar. 2002.
[9] X. Du, Y. Li, and D. Yao, “A Support Vector Machine Based Algorithm for Magnetic Resonance Image Segmentation,” Fourth International Conference on Natural Computation, 18-20 Oct. 2008.
[10] C. M. Wang, C. Y. Lin, J. A. Lin, and Z. H. Zheng, “Application of Extenics Approach in MRI Classification,” 18th IPPR Conference on Computer Vision, Graphics and Image Processing, Taipei, Taiwan, pp. 15-20, Aug. 2005.
[11] J. C. Su, C. M. Wang, S. C. Yang, and G. H. Chang, “An Extenics Approach to MRI Classification,” 2006 IEEE International Conference on Systems, Man, and Cybernetics, Taipei, Taiwan, pp. 562-567, Oct. 2006.
[12] W. E. Reddick, J. O. Glass, E. N. Cook, T. D. Elkin, and R. J. Deaton, “Automated Segmentation and Classification of Multispectral Magnetic Resonance Images of Brain using Artificial Neural Networks,” IEEE Trans. on Medical Imaging, vol. 16, no.6, pp. 911-918, Dec. 1997.
[13] J. S. Lin, R. M. Chen, and Y. M. Huang, “Medical Image Segmentation using Mean Field Annealing Network,” IEEE Conference on Image Processing, vol. 2, pp. 855-858, Oct. 1997.
[14] J. Alirezaie, M. E. Jernigan, and C. Nahmias, “Automatic Segmentation of Cerebral MR Images using Artificial Neural Networks,” IEEE Trans. on Nuclear Science, vol. 45, no. 4, pp. 2174-2182, Aug. 1998.
[15] J. S. Lin, K. S. Cheng, and C. W. Mao, “Multispectral Magnetic Resonance Images Segmentation using Fuzzy Hopfield Neural Network,” International Journal of Biomedical Computing, pp. 205-214, Aug. 1996.
[16] C. M. Wang, S. C. Yang, G. H. Chang, and J. A. Lin, “Extension Neural Network Approach in MRI Classification,” The 1st Intelligent Living Technology Conference, ILT 2006, Taichung, Taiwan, Jun. 2006.
[17] A. Mayer, and H. Greenspan, “An Adaptive Mean-Shift Framework for MRI Brain Segmentation,” IEEE Trans. on Medical Imaging, vol. 28, no. 8, pp. 1238-1250, Aug. 2009.
[18] A. Likas, N. Vlassis, and J. J. Verbeek, “The Global K-means Clustering Algorithm,” Pattern Recognition, vol. 36, no. 2, pp. 451-461, Feb. 2003.
[19] 楊建芳，李寬容，李三剛，李覃，楊康寧，何永仁，沈戊忠，韓念祖，蘇友吉，「磁振造影診斷學」，華榮出版社，1980。
[20] R. Adams, and L. Bischof, “Seeded Region Growing,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 16, no. 6, pp. 641-647, 1994.
[21] A. Mehnert, and P. Jackway, “An Improved Seeded Region Growing Algorithm,” Pattern Recognition Letters, vol. 18, no. 10, pp. 1065-1071, Oct. 1997.
[22] T. Pavlidis, and Y. T. Liow, “Integrating Region Growing and Edge Detection,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 12, no. 3, pp. 225-233, Mar. 1990.
[23] N. R. Pal, and S. K. Pal, “A review on Image Segmentation Techniques,” Pattern Recognition, vol. 26, no. 9, pp. 1277-1294, Sep. 1993.
[24] C. C. Chu, and J. K. Aggarwal, “The Integration of Image Segmentation Maps using Region and Edge Information,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 15, no. 12, pp. 1241-1252, Dec. 1993.
[25] A. Tremeau, and N. Borel, “A Region Growing and Merging Algorithm to Color Segmentation,” Pattern Recognition, vol. 30, no. 7, pp. 1191-1203, Jul. 1997.
[26] S. A. Hojjatoleslami, and J. Kittler, “Region Growing: A New Approach,” IEEE Trans. on Image Processing, vol. 7, no. 7, pp. 1079-1084, Jul. 1998.
[27] J. Fan, D. K. Y. Yau, A. K. Elmagarmid, and W. G. Aref, “Automatic Image Segmentation by Integrating Color-edge Extraction and Seeded Region Growing,” IEEE Transactions on Image Processing, vol. 10, no. 10, pp. 1454-1466, Oct. 2001.
[28] F. Y. Shih, and S. Cheng, “Automatic Seeded Region Growing for Color Image Segmentation,” Image and Vision Computing, vol. 23, no. 10, pp. 877-886, Sep. 2005.
[29] H. Gudbjartsson, and S. Patz, “The Rician Distribution of Noisy MRI Data,” Magn Reson Med. vol. 34, no. 6, pp. 910-914, Dec. 1995.
[30] C. E. Metz, “ROC Methodology in Radiological Imaging,” Invest Radiol, vol. 21, no. 9, pp. 720-723, Sep. 1986.