本論文主要目的是建立一套針對二維髖骨以及髖關節X光前後照影像偵測與評估髖關節脫臼及病變的系統，並且依據醫師專業知識建立模型以提供客觀的臨床評估數據。我們希望能藉此自動判讀分析系統輔助醫生縮短診斷流程並協助做出正確而迅速的治療或手術。
由於在影像中髖部隨著性別、年齡、體型等有很大的變異性存在；因此如果用一個固定的模型來套到每一張影像上面應該沒辦法達到絕佳的效果，然而它們又有著大致相同的形狀，因此如果能套用在一個可變的模型而這個模型又能兼顧觀察它們大致的形狀，那將能達到最佳的結果。動態形狀模型（Active Shape Models）是一個有彈性可變化又能吸取影像資訊的一個方法，因此我們將採用這個方法來進行初步的影像分割。
在解剖學上股骨、髂骨以及閉孔三個部分屬獨立不互相影響之部位。所以本論文將這些部位分開做以下獨立處理：(1)先由訓練樣本利用動態形狀模型建立出三個模型。(2)再透過地標點附近的灰階值找出正確的骨頭邊界。(3)接著評估Simon’s line 及Shenton’s line。 (4)從X-光影像中擷取出輪廓得到地標點之後，經由曲線擬合可以計算地標點離Simon’s line 及Shenton’s line 的距離最小平方誤差(5)並從誤差中自動找出可靠的臨界值以判別脫臼與否。

The goal of this thesis is to construct a system which analyzes the AP view X-ray images for hip and hip joint, and the system can also evaluate the extent of hip dislocation. The developed model of the proposed system is based on the doctor’s professional knowledge and provides objective parameters for clinical applications. We hope the proposed system can help doctor in decreasing the diagnostic time while improving the outcomes of therapy or surgery.
According to the sex, age, and body type, there exist many differences in hip images. Therefore, using a fixed model to fit every image does not work well. As the hips usually have similar shapes, an adaptive model which satisfies the general shape of hip may result in a better result. Active Shape Model (ASM) which can transform to accommodate the variable image information is such an elastic method. We adopt this method as the initial shape in segmenting the hip images.
The femur、ilium and obturator foramen belong to independent anatomical structures. This thesis segments the above structures using the following procedures: (1) Using active shape model to construct three models from the training pattern. (2) Refine the bone contour by using the nearby gray levels of landmarks. (3) Estimating the Simon’s line and Shenton’s line, respectively. (4) Based on the extracted contour from X-ray image, we can get the landmarks and calculate the least square error distance from the landmark to Simon’s line, or to the Shenton’s line. (5) Automatically calculating the error distance as the basis in diagnosing the dislocation of hip structure.

[1] A. Blake and M. Isard,“Active Contours.” Springer, 1998.

[2] F. L. Bookstein, “Landmark methods for forms without landmarks: Localizing group differences in outline shape.” Medical Image Analysis, 1(3):225-244, 1997.

[3] T. F. Cootes and C. J. Taylor,“Active shape models – ’smart snakes’.” In Proc. British Machine Vision Conf., BMVC92, pages 266–275, 1992.

[4] T. F. Cootes, A. Hill, C. J. Taylor and J. Haslam, "The Use of Active Shape Models for Locating Structures in Medical Images." Image and Vision Computing, Vol. 12, No. 6, pp.355-366, 1994.

[5] T. F. Cootes, C. J. Taylor, D. H. Cooper, and J. Graham, “Active shape models – their training and application.” Computer Vision and Image Understanding, 61(1):38–59, 1995.

[6] T. F. Cootes, G. J. Edwards, and C. J. Taylor, “Active appearance models.” In Proc. European Conf. on Computer Vision, volume 2, pages 484–498. Springer, 1998.

[7] T. F. Cootes, G. J. Edwards, and C. J. Taylor, “ Active appearance models.” IEEE Trans. on Pattern Analysis and Machine Intelligence, 23(6):681–685, 2001.

[8] M. Donna , K. F. Elliot, R. Derek, Ney, et al. “Acetabular and pelvis fractures in the pediatric patient value of two-and three-dimensional image.”, Journal of Pediatric Orthopaedics, 1992,12(5):621-625.

[9]T. F. Cootes G. J. Edwards, and C. J. Taylor, “ Interpreting face images using active appearance models.” In Proc. 3rd IEEE Int. Conf. on Automatic Face and Gesture Recognition, pages 300–5. IEEE Comput. Soc, 1998.

[10] I. L. Dryden and K. V. Mardia, “Statistical shape analysis.” John Wiley& Sons, 1998.

[11]R. Fisker, “Making Deformable Template Models Operational.”, PhD thesis, Department of Mathematical Modelling, Technical University of Denmark, Lyngby, 2000.

[12] C. Goodall,“ Procrustes methods in the statistical analysis of shape. “ Jour. Royal Statistical Society, Series B, 53:285–339, 1991.

[13] J. C. Gower, “ Generalized Procrustes analysis.” Psychometrika, 40:33–50, 1975.

[14] J. Isidoro and S. Sclaroff, “Active voodoo dolls: a vision based input device for nonrigid control.” In Proc. Computer Animation ’98, pages 137–143. IEEE Comput. Soc, 1998.

[15] A. K. Jain, Y. Zhong, and M.-P. Dubuisson-Jolly, “Deformable template models: A review.” Signal Processing, 71(2):109–129, 1998.

[16] M. J. Jones and T. Poggio, “Multidimensional morphable models: a framework for representing and matching object classes.” International Journal of Computer Vision, 29(2):107–31, 1998.

[17] M. Kass, A. Witkin, and D. Terzopoulos, “Snakes: Active contour models.” Int. Jour. of Computer Vision, 8(2):321–331, 1988.

[18] Michael J. Black and Allan D. Jepson. Eigentracking: Robust matching and tracking of articulated objects using a view-based representation. International Journal of Computer Vision, 26(1):63–84, 1998.

[19] S. Sclaroff and A. P. Pentl, “Modal matching for correspondence and recognition.” IEEE Transactions on Pattern Analysis and Machine Intelligence, 17(7):545–61, 1995.

[20] S. Sclaroff and J. Isidoro, “Active blobs.” Proc. of the Int. Conf. on Comput. Vision, pages 1146–1153, 1998.

[21] T. McInerney and D. Terzopoulos, “Deformable models in medical image analysis: a survey.” Medical Image Analysis, 2(1):91–108, 1996.
[22] B. M. Michael, B. M. Stephen, “Use of computed tomogtr-aphic reconstruction in planning osteotomies of the hip.” Clinical Orthopaedics And Related Research, 27 (1):154-159, 1992.