腦下垂體微小腺瘤在外觀上沒有明顯變化，診斷上依賴磁振造影動態掃描，在注射顯影劑前與注射顯影劑後一分鐘、兩分鐘、三分鐘和四分鐘共五個時間點的相同位置拍攝影像，觀察腦下垂體在不同時序的顯影變化。由於腦下垂體體積不大，且顯影訊號變化強度不明顯，造成臨床醫師判讀困擾。本論文主要目的是希望藉由影像處理的技術，先切割出腦下垂體，進而分析動態顯影曲線，在原始影像上標記分析結果，以協助臨床醫師診斷判讀病灶。
我們的方法包含腦下垂體影像分割和顯影曲線分析兩大部分。在腦下垂體影像分割部分，我們利用考量邊方向和強度的圓形霍氏轉換，找到影像中的內頸動脈。根據內頸動脈位置圈選出矩形區域，再使用 Otsu’s 方法得到門檻值將影像二值化，接著使用連結區塊運算並選取最大區塊，將五個時序影像的最大區塊做交集運算，切割出腦下垂體。在曲線分析部分，根據臨床診斷微小腺瘤的方法，制定出分群準則，將腦下垂體每個像素的顯影曲線分為七類，再依據臨床診斷經驗分為正常組織、疑似微小腺瘤和確定微小腺瘤三類，將結果標記回原圖，可提供臨床醫師異常曲線出現位置，做為診斷參考。

MRI dynamic contrast-enhanced technical acquisition allows evaluation of temporal changes in signal intensity within sella structures, it is useful on differentiating normal pituitary gland and microadenoma. Enhancement pattern of adenoma and normal pituitary gland shows as a time intensity curve. Signal intensity increases more rapidly and to a greater degree in normal pituitary gland than in microadenoma. However, pituitary gland volume is small and enhanced signal is not obvious. The proposed system of this paper is to analyze the time intensity curve of first segment of the pituitary gland of the MRI images through of image processing technology.
The proposed system consists of two main stages - pituitary gland segmentation and time intensity curve analyzation. In pituitary gland segmentation, we detect internal carotid artery using Hough transform and select a bounding box based on the location of the internal carotid artery. Binary image is obtained Otsu’s method within the bounding box and the largest connected component is selected as the location of pituitary gland. In time intensity curve analyzation , we cluster time intensity curves into seven types according to clinical diagnosis.
Finally, we provide cluster assignment maps based on the clustering results as a diagnostic aid.

[1]Ballard, D. H., "Generalizing the Hough transform to detect arbitrary shapes," Pattern Recognition, vol.13, no.2, pp.111-122, 1981.
[2]Berry, M., S. Suri, V. Chowdhury, and S. Mukhopadhyay,, eds., Diagnostic Radiology Neuroradiology Including Head and Neck Imaging, Jaypee Brothers Medical Publishers, New Delhi, 2006.
[3]Davies, E. R., "A new framework for analyzing the properties of the generalized Hough transform," Pattern Recognition, vol.6, no.1, pp.1-7, 1987.
[4]Duda, R. O. and R. E. Hart, "Use of the Hough transform to detect lines and curves in pictures," Communications of the ACM, vol.15, no.1, pp.11-15, 1972.
[5]Ertaş, G., H. O. Gülçür, and M. Tunaci, "An interactive dynamic analysis and decision support software for MR mammography, " Computerized Medical Imaging and Graphics, vol.32, no.4, pp.248-293, 2008.
[6]Gonzalez, R. C. and R. E. Woods, eds., Digital Image Processing, 3rd edition, Prentice Hall, New Jersey, 2009.
[7]Hakalahti, H., D. Harwood, and L. S. Davis, "Two dimensional object recognition by matching local properties of contour point," Pattern Recognition Letters, vol.2, no.4, pp.227-234, 1984.
[8]Hough, P. V. C., "Method and Means for Recognizing Complex Patterns," US Patent, December 18, 1962.
[9]Illingworth, J. and J. Kittler, "A survey of the hough transform," Computer Vision, Graphics and Image Processing, vol.44, no.1, pp.87-116, 1988.
[10]Indrajit, I. K., N. Chidambaranathan, K. Sundar, and I. Ahmed, "Value of dynamic MRI imaging in pituitary adenomas," Neuroradiology, vol.11, no.4, pp.185-190, 2001.
[11]Kimme, C., D. Ballard, and J. Sklansky, "Finding circles by an array of accumulators," Communications of the ACM, vol.18, no.2, pp.120-122, 1975.
[12]Kuhl C.K., P., S. Mielcareck, C. Klaschik, E. Leutner, J. Wardelmann, and H. H. S. Gieseke, "Dynamic breast MR imaging:are signal Intensity time course data useful for differential diagnosis of enhancing lesions?," Radiology, vol.211, pp.101-110, 1999.
[13]Leinsinger, G., T. Schlossbauer, M. Scherr, O. Lange, M. Reiser, and A. Wismüller, "Cluster analysis of signal-intensity time course in dynamic breast MRI: does unsupervised vector quantization help to evaluate small mammographic lesions?," European Radiology, vol.16, no.5, pp.1138-1146, 2006.
[14]Marieb, E. N., J. Mallatt, and P. B. Wilhelm, eds., Human Anatomy, 5th edition, Benjamin Cummings, 2007.
[15]Meyer-Baese, A., T. Schlossbauer, O. Lange, and A. Wismueller, "Small lesions evaluation based on unsupervised cluster analysis of signal-intensity time courses in dynamic breast MRI," International Journal of Biomedical Imaging, vol.2009, no.326924, 2009.
[16]Moring I. and Hakalahti H., "Scale independent method for object recognition," presented at the Scandinavian Conference on Image Processing., Trondheim, 1985.
[17]Namimoto, T., S. Morishita, R. Saitoh, J. Kudoh, Y. Yamashita, and M. Takahashi, "The value of dynamic MR imaging for hypointensity lesions of the peripheral zone of the prostate," Computerized Medical Imaging and Graphics, vol.22, no.3, pp.239-245, 1998.
[18]Sobel, I. E., Camera Models and Machine Perception, Ph.D. dissertation, Department of Computer Science, Stanford University, Palo Alto, California, 1970.
[19]Wismüller, A., O. Lange, D. R. Dersch, G. L. Leinsinger, K. Hahn, B. Pütz, and D. Auer, "Cluster Analysis of Biomedical Image Time-Series," International Journal of Computer Vision, vol.46, no.2, pp.103-128, 2002.
[20]Yuh, W. T. C., D. J. Fisher, H. D. Nguyen, E. T. Tali, F. Gao, T. M. Simonson, and J. A. Schlechte, "Sequential MR enhancement pattern in normal pituitary gland and in pituitary adenoma," American Journal of Neuroradiology, vol.15, no.1, pp.101-108, 1994