醫學影像辨識是一件頗為複雜的工作，必需要融合影像處理、醫學知識等多方面的知識才能加以完成。在眾多的醫學影像當中，磁核共振影像對器官病變的部份以及組織，能在影像能較明顯的反應出來，但因為高頻線圈的影響，磁核共振影像的灰階值範圍並不固定，因此也加深辨識的困難度。
本論文所設計實作的是針對在磁核共振影像，進行腦部以及脊椎器官的影像辨識。
在腦部方面，我們採用自動化區域成長方法，依照影像中腦部區域的特性，選取種子點，進行區域成長，再經由腦部的外觀及位置等特徵條件判斷，將腦部取出。在脊椎方面，採用半自動化區域成長的方式，由使用者在影像中框出一個在脊椎中的矩形，用使用者所定義的矩形取出脊椎在灰階值的特性，取出種子點，進行區域成長，經由脊椎的外觀及位置上特徵的處理，取出每個脊椎的部份，再進行改善的機制。

Medical image segmentation is a complex task that requires the integration of knowledge acquired from different domains such as medicine, image processing. In kind of medical images, the MRI has more obvious response for the organ pathological changes and tissue. Because of the high frequency coil effect, the gray value range is not steady in the MRI. So, the segmentation would be more difficult.
We design and implement at MR image, proceeds the brain and spine image segmentation.
In the brain, we use the automatic region growing method, according the feature of brain in the image to pick brain seeds, and then proceed the region growing, via, use the morphologic and position of the brain to judge the brain regions. In the spine, we use the half-automatic region growing method, defining a rectangle inside the spine in the image by the user. Using the rectangle to extract the gray value feature of spine, pick spine seeds, and then proceeds the region growing. Using the morphologic and position feature process of the spine to judge the spine regions, judging each spine region, to do the refine mechanism.

【PAPER】
[P1] M. Henkelman, “Measurement of signal intensities in the presence of noise in MR images,” Med. Phys., vol. 12, no. 2, pp. 232–233, 1985.
[P2] M. Stella Atkins and Blair T. Mackiewich, Fully Automatic Segmentation of the Brain in MRI, IEEE Transactions ON Medical Image, vol.17, no.1, FEBRUARY 1998
[P3] Laurence Germond, Michel Dojat , C. Taylor ,C. Garbay ,A cooperative framework for segmentation of MRI brain scans, Artificial Intelligence in Medicine 20 (2000) 77–93
[P4] YI-WEI YU and JUNG-HUA WANG, Image segmentation based on region growing and edge detection
[P5] Jacqueline Le Moigne and James C. Tilton, Refining image segmentation by integration of edge and region data , IEEE Transactions ON Geoscience and remote sensing vol.33 NO.3 MAY 1995

【BOOK】
[B1] Milan Sonka, Vaclav Hlavac, Rpger Boyle. Image Provessing, Analysis, and Machine Vision. 1999
[B2] 連國珍, 數位影像處理, 儒林圖書有限公司 1992

【WEB SITE】
[W1]http://www.resacorp.com/Rayleigh.htm