在放射線治療計畫中，核磁共振造影是最為廣泛使用的取像技術之ㄧ。許多運動員一旦在比賽中受傷都會去照核磁共振造影。醫生們藉由觀察核磁共振造影影像可以了解病人們關節、器官、或是肌肉附近的細節情況。
一直以來，人們因為老化或關節退化而受苦於前膝痛。醫生們作了多項的測試試圖去找出引起膝痛的原因。傳統上，靜態的核磁共振造影影像常被拿來觀察膝蓋附近軟組織黏附在大腿骨、髕骨及小腿骨的狀況。最近，動態的觀察技術也發展了。觀察組織的動態變化更可以幫助找出膝痛發生的原因。在我們的研究當中，我們觀察大腿骨與髕骨之間在膝蓋彎曲過程中的互動情況。為了量化觀察到的互動情形，醫生在大腿骨及髕骨的邊緣定義了幾個點，這些點決定了兩個定義的角度的計算。收集從正常膝蓋計算出來的角度的變化曲線，我們可以建構出一個正常範圍。從不正常膝蓋計算出來的角度變化曲線將會超出此正常範圍。藉由統計這些計算出來的角度變化，我們可以協助醫師檢查病人的膝蓋狀況。但是手動去標記這些定義點及計算角度是很耗費時間的。在本論文中，我們提出了一個系統整合了類神經網路技術與等階集合演化演算法去分割出影像中的大腿骨與髕骨、標記出定義點、並計算出相關角度。

In radiography treatment planning, MR image is one of the most widely-used radiographic techniques. Many athletes would take MR images once they get hurt in the game. Doctors can obtain detail information around the patient’s joints, organs, or muscle by observing the MR images.
All the time, people suffer from the anterior knee pain due to aging or degeneracy. Doctors do numbers of examinations hoping to find out the possible reasons causing the knee pain. In tradition, static MR images are often used to observe the situation of the soft tissue adhering to the femur, patella, and tibia. Recently, kinetic observation techniques are also developed. Observing the kinetic changes of the tissues can help even more in perceiving the reason why pain happens. In our research, we observe the interaction between the femur and patella during the bending process of the knee. To quantify the observed interaction, doctors set some landmark points on the boundaries of the femur and patella, which determine the calculation of two defined angles. The angle variation curves collected from normal knees can construct a normal range. The variation curve comes from abnormal knee exceeds the normal range. By the statistics of the variation of these angles calculated, we can help the doctor examine patient’s knee condition. But it takes lots of time to locate the landmarks and to calculate the corresponding angles manually. Here in this thesis, we propose a system combining neural network techniques and level set evolution algorithm to segment the femur and patella, locate landmarks and calculate the angles.

References

[1] Kuo-Sheng Cheng, Jzau-Sheng Lin, and Chi-Wu Mao. “The application of competitive hopfield neural network to medical image segmentation”, IEEE Transactions on Medical Image, vol.15, no.4, pp. 560-567, AUGUST 1996.

[2] Wilburn E. Reddick, John O.Glass, Edwin N.Cook, T. David Elkin, and Russell J. Deaton. “Automated segmentation and classification of multispectral magnetic resonance images of brain using artificial neural networks”, IEEE Transactions on Medical Image, vol.16, no. 6, pp. 911-918, 1997.

[3] Kazuhito Sato, Sakura Kadowaki, Hirokazu Madokoro, Masaki Ishi, and Atsushi Inugami. “Objective segmentation based on characteristics of single channel MR Images”, Proceedings of International Joint Conference of Neural Networks, pp. 1216-1221, 2007

[4] Chien-Cheng Lee, Pau-Choo Chung, and Hong-Ming Tsai. “Identifying multiple abdominal organs from CT image series using A multimodule contextual neural network and spatial fuzzy rules”, IEEE Transactions on Information Technology in Biomedicine, vol. 7, no. 3, pp.208-217, 2003.

[5] Chin-Tu Chen, Eric Chen-Kuo Tsao, and Wei-Chung Lin.”Medical image segmentation by a constraint satisfaction neural network”, IEEE Transactions on Nuclear Science, vol. 38, no. 2, pp.678-686, 1991.

[6] Yi-Nung Chung, Chen-Nan Chou, Haw-Chang Lan, Wen-Hsin Ho, “Kinematic patellar tracking from MR images for knee pain analysis” Elsevier Computers in Biology and Medicine, vol. 37, pp. 1653-1659, 2007.

[7] S.Osher, J.A. Sethian, “Fronts propagating with curvature-dependent speed: algorithms based on Hamilton-Jacobi formulations”, J. Comp. Phs., vol. 79, pp.12-49, 1988.

[8] Chunming Li, Chenyang Xu, Changfeng Gui, and Martin D. Fox. “Level set evolution without re-initialization: A new variational formulation”, IEEE Computer Society Conference on CVPR, vol. 1, pp.430-436, 2005.

[9] Fatih Kurugollu, B. Sankur, A. Emre Harmanci. “Image segmentation by relaxation using constraint satisfaction neural network”, Elsevier Science B.V. Trans. on Image and Vision Computing, vol. 20, pp. 483-497, 2002.

[10] T. Chan, and L. Vese, “Active contours without edges”, IEEE Trans. on Image Processing, vol.10, no. 2, pp.266-277, 2001.

[11] Pierrick Bourgeat, Sebastien Ourselin, Peter Stanwell, and Saad Ramadan, “Texture-based segmentation of the knee bones in MRI using phase information”, IEEE International Symposium on Biomedical Imaging: Nano to Macro. pp.742-745, 2006.

[12] Anil K. Jain, Farshid Farrokhnia, “Unsupervised texture segmentation using Gabor filters”, Elsevier Science B.V. Trans. on Pattern Recognition. vol. 24, pp.1167-1186, 1990.

[13] P. Campolucci, F. Capparelli, S. Guarnieri, F. Piazza and A. Uncini, “Neural networks with adaptive spline activation function”, 8th Mediterranean Electrotechnical Conference, vol 3. pp.1442-1445, 1996.

[14] Gregoire Allaire, Francois Jouve, Anca-Maria Toader, “Structural optimization using sensitivity analysis and a level-set method”, Elsevier B.V. Journal of Computational Physics, vol.194, pp.363-393, 2003

[15] Nikos Paragios, and Rachid Deriche, “Geodesic active regions and level set methods for supervised texture segmentation”, SpringerLink IJCV, vol. 46, pp.223-247, 2002.

[16] Stanley Osher, and Ronald P. Fedkiw, “Level set methods: an overview and some recent results”, Journal of Computational Physics, pp.169, 463, 2001.

[17] Jianping Fan, David. K. Y. Yau, “Automatic image segmentation by integrating color-edge extraction and seeded region growing”, IEEE Trans. on Image Processing, vol. 10, pp.1454-1466, 2001.

[18] T. Pavlidis, “Integrating region growing and edge detection”, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 12, pp.225-233, 1990.

[19] Chien-Chau Chu, and J. K. Aggarwal, “The integration of image segmentation maps using region and edge information”, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 15, pp.1241-1252, 1993.

[20] Frank G. Shellock, Jerrold H. Mink, “Patellar tracking abnormalities: Clinical experience with kinematic MR imaging in 130 patients”, NCBI Radiological Society of North America, pp. 799-804, 1989.

[21] Joachim Brossmann, Claus Muhle, “Patellar tracking patterns during active and passive knee extension: Evaluation with motion-triggered cine MR imaging”, NCBI Radiological Society of North America, pp. 205-212, 1993.

[22] A.H. Mir, M. Hanmandlu, S.N. Tandon, “Description of shapes in CT images. The usefulness of time-series modeling techniques for identifying organs”, IEEE Trans. on Engineering in Medicine and Biology Magazine, vol. 18, pp.79-84, 1999.

[23] Ning Xu, Narendra Ahuja, and Ravi Bansal “Object segmentation using graph cuts based active contours”, Elsevier Inc. Computer Vision and Image Understanding, vol.107, pp. 210-224, 2007.

[24] Michael Kass, Andrew Witkin and Demetri Terzopoulos, “Snakes: active contour models”, SpringerLink IJCV, vol. 1, pp.321-331, 1988.

[25] Bram van Ginneken, Alejandro F. Frangi, Joes J. Staal, Bart M. ter Haar Romeny, and Max A. Viergever, “Active shape model segmentation with optimal features”, IEEE Trans. on Medical Imaging, vol. 21, pp.924-933, 2002.

[26] Mikael Rousson and Nikos Paragios, “Shape priors for level set representations”, SpringerLink, Lecture Notes in Computer Science, vol. 2351, pp.416-418, 2002.