現今的手術導引系統多半皆使用單一的影像如電腦斷層影像(CT)、核磁共振影像(MRI)或C-arm X光影像，對於微創手術（Minimal Invasive Surgery）言，若能以術前的電腦斷層影像或核磁共振影像作三維的手術規劃，再以術中的二維C-am影像進行手術定位，則微創手術將更精準與可靠。本研究發展一套能夠整合C-arm影像及三維電腦斷層影像的方位校準方法，此方法不同於一般使用CT影像的手術導引系統，只需於術中拍攝兩張C-arm影像，即可完成病患與CT影像間的方位校準，以應用在微創手術。
首先將C-arm影像扭正並計算其發射源，然後選取C-arm影像與CT影像的相同特徵點，利用特徵點的座標完成病患座標系與CT模型座標系的初始方位校準，之後以濺射成像法(Splat Rendering)產生模擬C-arm影像的數位重建X-ray影像（Digital Reconstructed Radiograph, DRR），再以最佳化方法搜尋最佳的方位校準矩陣，使DRR影像與C-arm影像達到最大相似度。本研究以灰階基礎的影像相似度量測(Intensity Based Similarity Measure)結合特徵基礎影像相似度量測（Feature Based Similarity Measure）中的輪廓方法進行影像比對量測，並使用基因演算法（Genetic Algorithm）疊代運算完成方位校準的最佳化。實驗使用脊椎塑膠切骨模型（White Plastic Saw Bone），結果顯示DRR影像輪廓的平均定位誤差為0.26mm，平均角度誤差為0.78°；C-arm影像輪廓的平均定位誤差為0.53mm，平均角度誤差為1.45°；而在針對椎體(Vertebral Body)的特徵區域圈選實驗上，平均位移誤差及平均角度誤差分別為1.09mm與3.01度。

Most image-guided surgical navigation systems use single image modality such as computed tomography, magnetic resonance imaging, or C-arm X-ray imaging. It will enhance the precision and reliability of minimally invasive surgery if there are preoperative CT or MRI images for surgical planning and intraoperative C-arm images for surgical positioning. This research develops a 2D-3D registration method for the mapping between preoperative CT images and intraoperative C-arm images. The approach is different to that of CT image based navigation system. It only needs two intraoperative C-arm images to complete the registration and can be applied to minimally invasive surgery.
First, the C-arm images are calibrated and the focus point of X-ray is determined. Then, select three identical characteristic points in C-arm images and CT images to carry out the initial registration between the patient and its CT reconstructed model. After that, splat rendering is used to generate digital reconstructed radiographs (DRRs) to simulate C-arm images. Through the similarity measure of C-arm images and DRRs, the optimum transformation matrix can be found by applying optimization algorithms. This research combines the intensity based similarity measure method and the contour information of the feature based similarity measure method. Then, genetic algorithm is adopted to optimize the similarity rate and optimize the registration.
In the experiment, a white plastic saw bone model of spine is used. The average registration accuracy of location and orientation of DRR images are 0.26mm and 0.77° respectively. And the registration accuracy of location and orientation of C-arm contour information are 0.53mm and 1.45° respectively. Finally, the accuracy of location and orientation of a region of interest selection experiment stressed on vertebral body are 1.09mm and 3.01° respectively.

[1]Gebhard, F., Weidner, A., Liener, U. C., Stöckle, U., and Arand, M., “Navigation at the spine,” Injury, Int. J. Care Injured, Vol. 35, pp. S-A35-45, 2004.
[2]Livyatan, H., Yaniv, Z., Joskowicz, L., “Gradient-based 2-D/3-D rigid registration of fluoroscopic X-ray to CT,” IEEE Trans. Med. Imag., vol. 22, no. 11, pp. 1395-1406, Nov. 2003.
[3]BrainLAB, Germany, “BrainLabs Fluoro to CT Registration” http://www.brainlab.com/
[4]Levoy, M., “Efficient Ray Tracing of Volume Data,” ACM Trans. Graph., vol. 9 ,No. 3,1990.
[5]Russakoff, D.B., Rohlfing, T., Adler, J.R. Jr, and Maurer, C.R. Jr, “Intensity-based 2D-3D spine image registration incorporating a single fiducial marker,” Acad Radiol, Vol. 12, pp. 37-50, 2004.
[6]Weese, J., Gocke, R., Penney, G.P., Desmedt, P., Buzug, T.M., Schumann, H., “Fast voxel-based 2D/3D registration algorithm using a volume rendering method based on the shear-warp factorization,” SPIE International Symposium on Medical Imaging., San Diego, CA, USA, 1999.
[7]Ntasis, E., Maniatis, T.A., Nikita, K.S., “Fourier Volume Rendering for Real Time Preview of Digital Reconstructed Radiographs: A Web-Based Implementation,” Computerized Medical Imaging and Graphics26/1(1-8), 2002.
[8]Birkfellner, W., Seemann, R., Figl, M., Hummel, J., Ede, C., Homolka, P., Yang, X., Niederer, P., Bergmann, H., “Wobbled Splatting – a fast perspective volume rendering method for simulation of X-ray images from CT”, Phys Med Biol 50(9), pp. N73-N84, 2005.
[9]Joskowicz, L., Knaan, D., “How to achieve fast, accurate, and robust rigid registration between fluoroscopic X-ray and CT images,” CARS 2004., pp. 147-152. Chicago, IL, USA, 2004.
[10]Roth, M., Dötter, M., Burgkart, R., A. Schweikard, “Fast intensity-based fluoroscopy-to-CT registration using pattern search optimization,” CARS 2004., pp. 165-170, Chicago, IL, USA, 2004.
[11]Wein, W., “Intensity based rigid 2D-3D registration algorithms for radiation therapy”, Master’s thesis, Technische Universitat Munchen, German, Dec. 2003.
[12]Khamene, A., Bloch, P.,Wein, W., Svatos, M., Sauer, F., “Automatic registration of portal images and volumetric CT for patient positioning in radiation therapy,” Medical Image Analysis, 2006. , pp. 147-152.
[13]Penny, G. P., Weese, J., Little, J.A., Desmedt, P. , Hill, D.L.G., Hawkes, D. J. “A Comparison of Similarity Measures for Use in 2-D–3-D Medical Image Registration.” IEEE Trans. Med. Imag., vol.17, no.4, pp. 586-595,Aug. 1998.
[14]Mahfouz, M.R., Hoff, W.A., Komistek R.D., et al. “Effect of segmentation errors on 3D-to-2D registration of implant models in X-ray images”J Biomech(38),2005., pp.229-39.
[15]Iwashita, Y., Kurazume, R., Hara, K., Hasegawa T., “Fast Alignment of 3D Geometrical Models and 2D Color Images using 2D Distance Maps,” Proc. The 5th International Conference on 3-D Digital Imaging and Modeling (3DIM), pp.164-171, 2005.
[16]Cyr C.M., Kamal, A. F., Sebastian, T.B., Kimia, B.B., “2D-3D registration based on shape matching”, Proc. IEEE Workshop on Mathematical Methods in Biomedical Image Analysis., pp. 198-203, 2000.
[17]Morrison, D.F., “Applied Linear Statistical Methods”, Prentice-Hill, pp 57-72,1983.
[18]Strehl, A., Ghost, J., “Cluster Ensembles - A Knowledge Reuse Framework for Combining Multiple Partitions,” Journal of Machine Learning Research (3): pp. 583-617,2002.
[19]Accuray Inc., “CyberKnife® Robotic Radiosurgery System” http://www.accuray.com/
[20]Yang, X., Birkfellner, W., Niederer, P., “A similarity measure based on Tchebichef moments for 2D/3D medical image registration.” CARS 2004., pp. 153-158, Chicago, IL, USA, 2004.
[21]Deguchi, K., “Optimal Motion Control for Image-Based Visual Servoing by Decupling Translation and Rotation.” Proc. IEEE/RSJ Intl. Conference on Intelligent Robots and System, pp. 706-711, Victoria, B.C., Canada,1998.
[22]Gueziec, A., Kazanzides P., “Anatomy Based Registration of CT-Scan and Intraoperative X-Ray Image for Guiding a Surgical Robot.” IEEE Trans. Med. Imag., vol.17, no. 5, pp. 715-727,Oct. 1998.
[23]Bansal, R., Staib, L.H. and Chen, Z., et al., “Entropy-based, multiple-portal-to-3D CT registration for prostate radiotherapy using iteratively estimated segmentation,” Proc. MICCAI ‘99, vol. 1679 of Lecture Notes in Computer Science, pp. 567–578, Cambridge, UK, Sept. 1999.
[24]夏笙, “應用於椎莖骨釘植入的手術導引系統”, 碩士論文, 中央大學機械工程研究所, 2005.
[25]楊遠祥, “應用於股骨轉子間骨折回復手術之C-arm based 手術導引系統”, 碩士論文, 中央大學機械工程研究所, 2005.
[26]戴君益, “C-arm影像與電腦斷層影像之方位校準方法”, 碩士論文, 中央大學機械工程研究所, 2007.
[27]蘇木春、張孝德“機器學習：類神經網路、模糊系統及基因演算法則”, 全華圖書科技股份有限公司, Ch9,1999.