現今流行的全膝關節置換術或單膝關節置換術都是切除大面積的骨組織，然而依據Widuchowski 的臨床統計，膝關節的病變範圍大約有90%是在2 平方公分之內。若針對病變區做局部置換手術，除了有傷口小、復原快的優點外，也保留了大部分的正常軟硬骨組織。近年來已有局部膝關節置換術通過FDA 認證，但這些植入物並非客製化品。若能在術前偵測出膝關節軟骨損傷區的表面型狀，則可以預製修補缺陷所需的客製化人工植入物，達到客製化局部膝關節置換的目標。
本研究使用核磁共振影像(MRI)來判讀膝關節軟骨的邊界。MRI 影像先經由影像處理和調整Window/Level 的方式，偵測出每張MRI 影像膝關節軟骨表面的邊界，再經由三維重建的技術，建立出膝關節軟骨表面的三維模型。為了確認所重建之三維模型的正確度，以六隻豬後腿的關節作實驗，並進行MRI 掃描與關節軟骨表面雷射掃描。由MRI 影像重建的三維模型再與雷射掃描所獲得的點群座標進行誤差比對，以確認由MRI 影像重建之關節軟骨曲面的精確度。
實驗結果發現，由MRI 重建之三維模型與由雷射掃描所得的點之間的整體平均誤差為0.438mm，而在易受損的內髁、滑車和外髁等部位，其平均誤差分別為0.326mm、0.253mm和0.31mm。而豬膝關節MRI 的像素大小為0.4297mm ´0.4297mm，影像間距為1mm，故局部平均誤差皆在1 個像素內。所以以MRI 影像評估膝關節受損區似乎可行，換句話說，可將此重建之三維模型作為局部膝關節修補用植入物的設計依據。

Currently popular total knee replacement or unicompartmental knee replacement has to resect large volume of bone tissue. However, according to the clinic statistics of Widuchowski, around 90 percent of knee joint defects are within 2cm2 in area. If the defect is treated by partial joint replacement, the advantages include smaller incision, quicker recovery, and less resection of normal tissue. In recent years, FDA had approved implants for partial knee replacement. However, these implants are not customer-made. If the surface shape of the defect area of knee cartilage can be detected and diagnosed prior to operation, the customer-made implant for replacing the defect can be prefabricated and enable the feasibility of using customer-made implants for partial knee replacement.
In this study, MRI (magnetic resonance imaging) images are used to diagnose the cartilage boundary of knee joints. Through image processing and adjustment of Window / Level of the images, the cartilage boundary of knee joints of each MRI image can be detected. Then, the three-dimensional model of knee cartilage can be reconstructed by using marching cube algorithm. In order to confirm the accuracy of the reconstructed cartilage 3D model, six pig hind leg joints are conducted to have both MRI scan and laser scan. The accuracy of 3D cartilage surface models reconstructed from MRI images is verified by the point data obtained from laser scan.
The experimental results showed that the overall average error between the MRI 3D model and the points obtained by laser scanning is 0.438mm. Moreover, the high-potential damage areas such as the medial condyle, trochlea, and the lateral condyle, their average errors are 0.326 mm , 0.253 mm and 0.31 mm, respectively. Compared to the MRI pixel size of 0.4297mm by 0.4297mm and image spacing of 1mm, the average errors are smaller than the pixel size. Therefore, it seems to be feasible to assess the knee damage area from MRI images. In other words, the reconstructed 3D model from MRI images can be used as a reference for the design of implants for partial knee arthroplasy.

[1]. Widuchowski, W., Widuchowski, J., and Trzaska, T., "Articular cartilage defects: Study of 25,124 knee arthroscopies.", The Knee, Vol. 14, pp. 177-182, Feb. 2007.
[2]. Hoemann, C. D., Hurtig, M., Rossomacha, E., et.al, "Chitosan-glycerol phosphate/blood implants improve hyaline cartilage repair in ovine microfracture defects.", The Journal of Bone and Joint Surgery, Vol. 87-A, pp. 2671–2686, Dec. 2005.
[3]. Sharpe, J. R., Ahmed, S. U., Fleetcroft, J. P., et.al, "The treatment of osteochondral lesions using a combination of autologous chondrocyte implantation and autograft.", The Journal of Bone and Joint Surgery, Vol. 87-B, pp. 730-735, May 2005.
[4]. Steinwachs, M., Guggi, T., and Kreuz, P., "Marrow stimulation techniques.", Injury, Vol. 39, pp. S26-S31, 2008.
[5]. Kuroda, R., Ishida, K., Matsumoto, T., et.al, "Treatment of a full-thickness articular cartilage defect in the femoral condyle of an athlete with autologous bone-marrow stromal cells.", Osteoarthritis and Cartilage, Vol. 15, pp. 226-231, 2007.
[6]. Kirker, C. A., Sickle, A. C. V., Steve, E. W., et al., "Safety of, and biological and functional response to, a novel metallic implant for the management of focal full-thickness cartilage defects: preliminary assessment in an animal model out to 1 year.", Journal of Orthopaedic Research, Vol. 24, No. 5, pp. 1095-1108, May. 2006.
[7]. Berg, V. B. C., Lecouvet, F. E., Poilvache, P., et.al, "Assessment of Knee cartilage in cadavers with dual-detector spiral CT arthrography and MR imaging.", Radiology, Vol. 222, pp. 430-436, 2002.
[8]. Daenen, B. R., Ferrara, M. A., Marcelis, S., et.al, "Evaluation of patellar cartilage surface lesions: comparison of CT arthrography and fat-suppressed FLASH 3D MR imaging.", European Radiology, Vol 8, pp. 981-985, 1998.
[9]. Kass, M., Witkin, A., and Terzopoulos, D., "Snakes: active contour models.", International Journal of Computer Vision, Vol. 1, pp. 321-331, 1988.
[10]. Horn, B. K. P. and Schunck, B. G., "Determining optical flow.", Artificial Intelligence, Vol. 17, pp. 185-204, 1981.
[11]. Xu, C. and Prince, J. L., "Gradient vector flow: a new External force for snakes.", IEEE Computer Society Conf., Computer Vision and Pattern Recognition, Vol. 17, No. 19, pp. 66–71, June 1997.
[12]. Xu, C. and Prince, J. L., "Snakes, shapes, and gradient vector flow.", IEEE transactions on image processing, Vol. 7, No. 3, pp. 359-369, March 1998.
[13]. Chi, Y., Cashman, P., and Kitney, R., "Automatic segmentation of cartilage in MR images using CDCG：chessboard directional compensated GVF snakes.", International Conference on MedicalInformation Visualisation (MediVis 06), IEEE Computer Society, pp. 45-50, 2006.
[14]. Lynch, J. A., Zaim, S., Zhao, C. G., et.al, "Automating measurement of subtle changes in articular cartilage from MRI of the knee by combining 3D image registration and segmentation.", Proceedings of SPIE, Vol. 4322, pp. 431-439, 2001.
[15]. Solloway, S., Hutchinson, C., Waterton, J., et.al, "The use of active shape models for making thickness measurements of articular cartilage from MR images.", Magnetic Resonance in Medicine, Vol. 37, No. 6, pp. 943–952, 1997.
[16]. Folkesson, J., Dam, E. B., Olsen, O. F., et.al, "Segmenting articular cartilage automatically using a voxel classification approach.", IEEE Transactions on Medical Image, Vol. 26, No. 1, 2007.
[17]. Lorensen, W. E. and Cline, H. E., "Marching cubes: a high resolution 3D surface construction algorithm.", Computer Graphics, Vol. 21, No. 4, pp. 163-169, 1987.
[18]. Besl, P. J. and Mckay, N. D., "A method for registration of 3-D shapes.", IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 14, pp. 239-256, 1992.
[19]. Grau, V., Mewes, A., Alcaiz, M., et.al, "Improved watershed transform for medical image segmentation using prior in-formation.", IEEE Trans. Med. Imag., Vol. 23, no. 4, pp. 447-458, April. 2004.
[20]. 蘇文, "膝關節局部表面重建研究.", 碩士論文, 中央大學機械工程研究所, 2008.
[21]. Tamez-Pena, J. G., Barbu-McInnis, M. and Totterman, S., "Knee cartilage extraction and bone-cartilage interface analysis from 3D MRI data sets.", Proceedings of the SPIE, Vol. 5370, pp. 1774-1784, 2004.
[22]. Huang, H. C., Chen, C.M., Wang, S. D., et.al, "A new noise reduction approach based on the slope facet model.", Applied Optics, Vol. 40, No. 29, pp. 5192-5205, Oct. 2001.
[23]. Zhang, D. S. and Kouri, D. J., "Varying weight trimmed mean filter for the restoration of impulse noise corrupted images.", in IEEE Int. Conf. Acoust. Speech Signal Process., pp. 137–140, 2005.
[24]. Dong, Y. and Xu, S., "A new directional weighted median filter for removal of random-valued impulse noise.", IEEE Signal Process. Lett., Vol. 14, No. 3, pp. 193–196, Mar. 2007.
[25]. Beghadane, A. and Khellaf, A., "A noise-filtering method using a local information measure.", IEEE Trans. Image Process, Vol. 6, No. 6, pp. 879-882, June 1997
[26]. Kumar, V. R. V., Manikandan, S., Vanathi, P. T.,et.al, "Adaptive window length recursive weighted median filter for removing impulse noise in images with details preservation.", Ecti transactions on electrical, Vol. 6, No. 1, Feb. 2008.
[27]. Thillou, C. M. and Mirmehdi, M., " Super-resolution text using the teager filter.", First International Workshop on Camera-Based Document Analysis and Recognition, pp. 10-16, Aug. 2005.
[28]. Mithoefer, K., Williams, R. J., Warren, R. F., et.al, "Chondral resurfacing of articular cartilage defects in the knee with the microfracture technique.", The Journal of Bone and Joint Surgery (American), Vol. 88, pp. 294-304, 2006.
[29]. Fitzgibbon, A. W., Pilu, M., and Fisher, R. B., "Direct least-squares fitting of ellipses.", IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 21, No. 5, pp. 476-480, May. 1999.
[30]. Canny, J., "A computational approach to edge detection.", IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 8, No. 6, Nov. 1986.
[31]. Taubin, G., "A signal processing approach to fair surface design.", Computer Graphics Proceedings, Vol. 29, pp. 351-358, Aug. 1995.
[32]. Cohen, Z. A., McCarthy, D. M., Kwak, S. D., et.al, "Knee cartilage topography, thickness, and contact areas from MRI: in-vitro calibration and in-vivo measurements.", Osteoarthritis and Cartilage, Vol. 7, No. 1, pp. 95-109, Jan. 1999.