磁振成像（MRI）檢驗是目前在臨床診斷上相當普遍的檢驗。在臨床應用上，MRI較CT在肌腱與肌肉等軟組織具有較高的分辨力。因此在臨床診斷時，MRI在顯示人體的器官組織與其之間的關係上，具有相當大的實用性。然而在傳統的影像顯示系統，MR序列影像通常直接由一組二維序列影像內插獲得，所以通常無法顯示出切面與切面之間器官組織的詳細變化。因此本論文提出一套方法，整合三個二維MR影像序列之資訊，進而重構出三維影像資料，並分割出在此三維影像資料中所欲觀察之組織。

本篇論文提出的方法主要包含兩部份：第一為如何由二維影像來建構出三維資料，第二為如何從建構出的三維資料中分割出欲觀察的物件，在第一部份，三維資料的重構上，由於MR序列影像通常由冠狀、矢面及軸向等三組不同的垂直方向來取像，因此在重構三維資料之前，必須先對此三組影像作對位處理。而本論文亦提出一套三組序列影像的對位程序，此對位程序主要分為兩個步驟：先將兩組序列影像作對位處理，再將第三組序列影像與之前對位完成的兩組序列影像作對位，而得到最後的對位結果，然後在三維資訊的重構上則可參考其三種不同方向的影像資訊，來建構出較為精確的三維體積資料。以往由二維影像資訊建構三維資訊時，通常僅由單一方向的影像序列來作為重構時的參考。本篇論文所提出的方法，則是以對位過後的三組資料來當作三維資料重構時的參考。故在精確度與三維影像品質上，經過實驗證明，均比傳統重構方法較為精確。

在第二部份，三維分割的方法上，我們利用最大事後機率估測法推導出在分割所欲最佳化之能量函式。此能量函式包含：高斯雜訊模型，形狀差異模型及馬可夫隨機場模型。高斯雜訊模型主要是控制分割結果中灰階值的分佈，形狀差異模型則是控制著分割物件的形狀，而馬可夫隨機場模型則是控制分割物件的平滑度。在分割的過程中，我們首先以基本的分割方法分割出一個初始的結果，然後將結果套入分割模型，將初始分割結果最佳化，進而得到最後的分割結果。最後將分割完成之物件，以OpenGL顯示。在影像分割的實驗中，我們討論三種不同的分割模型對分割結果有何影響。並將此套方法應用於肩部及腦部之資料，重構出肩部與腦部組織，結果發現本篇論文所提出的方法能夠精確的分割出在體積資料中的三維物件。

Magnetic resonance imaging (MRI) is a popular medical imaging modality widely used in clinical evaluation and treatment. The appearance of soft-tissue such as tendon and muscle in MR images is superior to the one of conventional computerized tomography (CT). As a result, it is most suitable to display all kinds of organs and their relationship in clinical diagnosis. However the multi-spectral MR images are usually acquired as a 2D image sequence along a given direction, it is difficult to reconstruct the volumetric information, e.g. volume or surface variation of organs, from a single image sequence. In this thesis, we proposed to use three 2D MR image sequences acquired from three orientations in clinical practice to construct more precisely the 3D organs, and then to segment the 3D volume of organs in this thesis.

The proposed technologies in this thesis are divided into two main parts. One is to reconstruct the 3D volume from multiple MR image sequences and the other is to segment the organs from the created 3D volume. In the process of 3D volume reconstruction, the MR image sequences are acquired in three different orientations that are coronal, axial and sagittal views. Before 3D volume reconstruction, we must at first align these image sequences. The proposed registration method consists of two stages. At stage one, we align the coronal and sagittal views, then we register the axial view to the aligned result from the first previous step to get the final result. After registration, we can use images of three different views as references to reconstruct the volume information. Unlike the conventional 3D volume reconstruction from single image sequence, our method compute the 3D interpolated data directly referring to three 2D MR image sequences. In the experiments, it shows that the proposed method is more accurate in volume reconstruction than the conventional ones.

As for the 3D organ segmentation, we adopt the Maximum a Posteriori estimation to derive an energy function for segmentation. The results are obtained by maximizing the energy function which is composed of the Gaussian noise model, shape difference model, and Markov Random Fields model. The Gaussian noise model represents the gray-value distribution of the segmented object. The shape difference model restricts the shape of segmented object. And the MRF model controls the smoothness of segmented object. The proposed method first obtains the initial model of the segmented object based on simple thresholding and then optimizes the energy to get the final segmentation result. Furthermore, the resulted segmented objects are displayed by using OpenGL library. In the experiments, the proposed segmentation algorithms are applied to MR shoulder image sequence and MR brain image sequence. We also discuss how the three models affect the segmentation results. The experimental results reveal that our proposed method can effectively interpolate and segment the 3D objects from multi-spectral MR image sequences.

[1] Y. Yang, A. Tannenbaum, and D. Giddens ”Knowledge-based 3D Segmentation And Reconstruction of Coronary Arteries Using CT Images, ” in Proc. Engineering in Medicine and Biology Society (EMBS). vol. 1, pp. 1664-1666, 2004.
[2] J. G.. Snel, H. W. Venema, and C. A. Grimbergen, ”Deformable Triangular Surfaces Using Fast 1-D Radial Lagrangian Dynamics-Segmentation of 3-D MR And CT Images of the Wrist,” IEEE Trans. Medical Imaging, vol.21, no. 8, pp. 888-903, Aug. 2002.
[3] R. A. Zoroofi, Y. Sato, T. Sasama, T. Nishii, N. Sugano, K. Yonenobu, H. Yoshikawa, T. Ochi, and S. Tamura, “Automated Segmentation of Acetabulum And Femoral Head From 3-D CT Images,” IEEE Trans. Information Technology in Biomedicine, vol. 7, no. 4, pp. 329-343, Dec. 2003.
[4] R. A. Zoroofi, Y. Sato, T. Sasama, T. Nishii, N. Sugano, H. Yoshikawa, and S. Tamura, “Automated Segmentation of Necrotic Femoral Head From 3-D MR Data,” Computerized Medical Imaging and Graphics (CMIG), vol. 28, no. 5, pp. 267-278, Jul. 2004.
[5] R. A. Zoroofi, Y. Sato, T. Nishii, N. Sugano, H. Yoshikawa, and S. Tamura, “Segmentation of Avascular Necrosis of The Femoral Head Using 3-D MR Data,” Computerized Medical Imaging and Graphics(CMIG), vol. 25, no.6, pp. 511 – 521, Nov.-Dec. 2001.
[6] Y. T. Chen and M. S. Wang, “Three-Dimensional Reconstruction and Fusion For Multi-modality Spinal Images,” Computerized Medical Imaging and Graphics, vol. 28, no. 1-2, pp. 21–31, Jan. 2004.
[7] Yan Kang, K. Engelke, and W. A. Kalender, “A New Accurate and Precise 3-D Segmentation Method for Skeletal Structures in Volumetric CT Data,” IEEE Trans. Medical Imaging, vol. 22, no. 5, pp. 586-598, May 2003.
[8] Yan Kang, K. Engelke, and W. A. Kalender, “Interactive 3D Editing Tools For Image Segmentation,” Medical Image Analysis, vol. 8, no. 1, pp 35-46, Mar. 2004.
[9] X. Zhao, F. Qi, H. Huang, and J. Zhan, “3-D deformable models for femur shape recovery from volumetric images,” in Proc. The Fourth International Conference on High Performance Computing in the Asia-Pacific Region, vol. 2, pp. 797-798, May 2000.
[10] A.B. Albu, D. Laurendeau, L. J. Hebert, H. Moffet, and C. Moisan, “Three-dimensional Reconstruction of the Bony Structures Involved in the Articular Complex of the Human Shoulder Using Shape-based Interpolation and Contour-based Extrapolation,” in Proc. Fourth International Conference on 3-D Digital Imaging and Modeling (3DIM), pp. 370-377, 2003.
[11] M.-E. Tremblay, A. B. Albu, L. Hebert, Denis Laurendeau, “Integrating Region and Edge Information For the Automatic Segmentation of Interventional Magnetic Resonance Images of the Shoulder Complex,” in Proc. First Canadian Conference on Computer and Robot Vision, pp. 279-286,May 2004.
[12] R. C. Gonzalez and R. E. Woods, “Digital Image Processing”. New York, Addison-Wesley, 1992.
[13] J. K. Udupa and R. J. Goncalves, “Imaging Transforms for visualizing surface and volumes,” J. Digital imaging, pp.213-236, 1993.
[14] M.R. Stytz and R.W. Parrott, “Using kriging for 3-D medical imaging,” Computerized Med. Imaging and Graphics, pp.421-442, 1993.
[15] http://www.environment.sa.gov.au/mapland/sicom/sicom/tp_scs.html.
[16] T. N. Tran, R. Wehrens and L. M.C. Buydens, “Clustering Multispectral Images: a Tutorial,” Chemometrics and Intelligent Laboratory Systems, vol. 77, no. 1-2, pp. 3-17, May 2005.
[17] A.K. Shackelford and C.H. Davis, “A Hierarchical Fuzzy Classification Approach For High-resolution Multispectral Data Over Urban Areas,” IEEE Trans. Geoscience and Remote Sensing, vol. 41, no. 9, pp.1920-1932, Sept. 2003.
[18] P. Paclík, R. P. W. Duin, G. M. P. van Kempen, and R. Kohlus, ”Segmentation of Multi-spectral Images Using The Combined Classifier Approach,” Image and Vision Computing, vol. 21, no. 6, pp. 473-482, 2003.
[19] A. Clément and B. Vigouroux, “Unsupervised Segmentation of Scenes Containing Vegetation (Forsythia) and Soil by Hierarchical Analysis of Bi-dimensional Histograms,” Pattern Recognition Letters, vol. 24, no. 12, pp. 1951-1957, 2003.
[20] N. Otsu, “A Threshold Selection Method From Gray Level Histograms,” IEEE Trans. Syst. Man. Cybern., vol. 9, pp. 62-66, 1979.
[21] M. Kass, A. Witkin, and D. Terzopolous, ”Snake: Active Contour Models,” Int. J. Comput. Vis. 1, pp. 321-331, 1988.
[22] S. Osher and J. A. Sethian, “Fronts Propagating With Curvature Dependent Speed: Algorithms Based on Hamilton-Jacobi Formlation,” Journal of Computational Physics, vol. 79, no. 1, pp. 12-49, Nov. 1988.
[23] T. F. Cootes, C. J. Taylor, D. H. Cooper, and J. Graham, “Active Shape Models: Their Training and Application,” Computer Vision and Image Understanding, vol. 61, no. 1, pp. 38-59, 1995.
[24] T. Zoller and J. M. Buhmann, “Shape Constrained Image Segmentation By Parametric Distributional Clustering,” in Proc. IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), vol. 1, pp. I-386 - I-393, 2004.
[25] http://www.bic.mni.mcgill.ca/brainweb/