近幾年來，內視鏡手術因具有手術傷口小、病人復原時間短暫和節省醫療成本等優點，而在臨床上被廣泛應用。在手術中，使用內視鏡來觀察腔室有如使用單一的眼睛來觀察物體，往往使醫生無法正確的判斷出腔室區域的大小以及其立體空間的關係。因此，本論文提出了一個內視鏡輔助系統，以解決這些問題。
　　此系統利用內視鏡影像序列重構出腔室的三維模型，並在手術工具上固定一三維追蹤器miniBIRD來更新該模型，並即時顯示手術工具之行進與方位資訊。此系統主要可分為四個部分：特徵點的擷取與追蹤、三維模型的建構、三維腔室模型的修正與更新以及手術工具模型的互動式顯示。
　　我們根據人體生理組織的形狀起伏和色彩變化來強化內視鏡影像的特徵，並藉由邊線特徵(edge feature)的擷取以及特徵點的運動限制來改進Kanade-Lucas-Tomasi (KLT)特徵追蹤演算法，以獲得多張影像的對應點。我們也改良分解法(factorization method)，由各影像的對應點以透視投影模型疊代逼近平行透視投影模型求出特徵點的三維座標，該三維模型將和實際腔室同樣大小。為了進一步的修正和更新腔室模型，醫師可以利用手術工具在特別關心的區域點選數個具代表性的特徵點，並將之帶入分解法中重新計算出一個修正過的三維模型。重複執行此步驟，我們將可獲得一逼近真實腔室的三維模型，以提供醫師相關的數據量測。
　　另一方面，因為手術工具上固定一三維追蹤器，我們可以即時的顯示出手術工具模型的位置。如此一來，腔室與手術工具間的空間關係將可被視覺化。手術工具的位置、行進速度和插入人體的角度等資訊將可提供醫師正確的手術預警。

　　Recently, the endoscopic minimally invasive surgery becomes popular than ever before because of its advantages in small incisions, faster recovery of the patients, and reduction of medical costs. However, seeing the images from endoscopes is similar to see scenes with only one naked eye. The surgeon usually cannot determine the size of organs or the distance between them exactly due to the lack of stereo visual perception.
　　In order to resolve this problem, we propose a computer-assisted system for endoscopic surgery in this thesis. This system reconstructs a 3-D model of human organs or tissues from an endoscopic images sequence. And a 3-D tracker called miniBIRD is attached to the surgical instrument to obtain the 3-D position of instrument and to update the reconstructed structure of human organs or tissues. This system is composed of four procedures: feature extraction and tracking, 3-D model reconstruction, 3-D model updating, and instrument model visualization.
　　According to the shape and texture of organs, we enhance all images by histogram equalization and intensity normalization. Moreover, the Kanade-Lucas-Tomasi (KLT) tracker is improved by extracting edge features and using the motion restriction mechanism. The tracked 2-D feature points by KLT tracker are then used to reconstruct the 3-D model by using the factorization algorithm based on the paraperspective projection. The factorization algorithm is modified by adopting 3-D points with known coordinates to build a 3-D model with real size. For refining the reconstructed 3-D structure, the surgeon can acquire additional 3-D positions from some characteristic feature points by miniBIRD and perform the improved factorization method again to generate a more accurate model. In addition, the miniBIRD is attached to the surgical instrument whose position is consequently tracked in real-time. Since the organ structure and surgical instrument are both in the same coordinate system, the distance between organs and instrument, the velocity of instrument, and the inserting angle of instrument can all be estimated to give surgeons a proper warning during the operation.

[1] T.Y. Lee, P.H. Lin, C.H. Lin, Y.N. Sun, and X.Z. Lin, “Interactive 3D Virtual Colonoscopy System,” IEEE Trans. Information Technology in Biomedicine, vol. 3, no. 2, pp. 139-150, June 1999.

[2] J.P. Helferty, C. Zhang, G. McLennan, and W.E. Higgins, “Videoendoscopic Distortion Correction and Its Application to Virtual Guidance of Endoscopy,”IEEE Trans. Medical Imaging, vol. 20, no. 7, pp. 605-617, July 2001.

[3] R. Shahidi, M.R. Bax, C.R. Maure, J.A. Johnson, E.P. Wilkinson, B. Wang, J.B. West, M.J. Citardi, K.H. Manwaring, and R. Khadem, “Implementation, Calibration and Accuracy Testing of an Image-Enhanced Endoscopy System,” IEEE Trans. on Medical Imaging, vol. 21, no. 12, pp. 1524-1535, December 2002.

[4] K.C. Patel, J.L. Duerk, Q. Zhang, Y.C. Chung, M. Williams, K. Kaczynski, M. Wendt, and J.S. Lewin, “Methods for Providing Probe Position and Temperature Information on MR Images During Interventional Procedures”, IEEE Trans. Medical Imaging, vol. 17, no. 5, pp. 794-802, October 1998.

[5] S.L. Hartmann and R.L. Galloway, “Depth-Buffer Targeting for Spatially Accurate 3-D Visualization of Medical Images”, IEEE Trans. Medical Imaging, vol. 19, no. 10, pp. 1024-1031, October 2000.

[6] D. Dey, D.G. Gobbi, P.J. Slomka, K.J.M. Surry, and T.M. Peters, “Automatic Fusion of Freehand Endoscopic Brain Images to Three-Dimensional Surfaces: Creating Stereoscopic Panoramas”, IEEE Trans. Medical Imaging, vol. 21, no. 1, pp. 23-30, January 2000.

[7] R. Hartley and A. Zisserman, Multiple View Geometry in Computer Vision, Cambridge University Press, 2000.

[8] M. Pollefeys, R. Koch and L.V. Gool, “Self-calibration and metric reconstruction in spite of varying and unknown internal camera parameters”, Proc. 6th Int’l Conf. Computer Vision (ICCV 98), pp. 90-95, January 1998.

[9] C. Tomasi and T. Kanade, “Factoring image sequences into shape and motion,”Proc. of the IEEE Workshop on Visual Motion, pp.21-28, October 1991.

[10] C. J. Poelman and T. Kanade, “A paraperspective factorization method for shape and motion recovery,” Carnegie Mellon University Tech. Rep. CMU-CS-93-219, 11 December, 1992.

[11] S. Christy and R. Horaud, “Euclidian shape and motion from multiple perspective views by affine iterations,” INRIA Tech. Rep. RR-2421, December 1994.

[12] K. Deguchi, T. Sasano, and H. Arai et al., “3-D shape reconstruction from endoscope image sequences by the factorization method,” IEICE Trans. Inf. & Syst., vol. E79-D, no. 9, pp.1329-1336, September 1996.

[13] S.M. Yamany, A.A. Farag, D. Tasman, and A.G. Farman, “A 3-D reconstruction system for the human jaw using a sequence of optical images,” IEEE Trans. Medical Imaging, vol. 19, no. 5, pp. 538-547, May 2000.

[14] R. Horaud, S. Christy, F. Dornaika, and B. Lamiroy, “Object pose: the link between weak perspective, paraperspective, and full perspective,” International Journal of Computer Vision, vol. 22, no. 2, pp. 173-189, March 1997.

[15] C. Tomasi and T. Kanade, “Detection and tracking of point features,” Carnegie Mellon University Tec. Rep. CMU-CS-91-132, April 1991.

[16] J. Shi and C. Tomasi, “Good features to track,” Proc. of the IEEE Conf. Computer Vision and Pattern Recognition, pp. 593-600, June 1994.

[17] T. Tommasini, A. Fusiello, and E. Trucco et al., “Making good features track better,” Proc. of the IEEE Conf. Computer Vision and Pattern Recognition, pp. 178-183, June 1998.

[18] H. Aanæs, R. Fisker, K. Åström, and J.M. Carstensen, “Robust Factorization,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 9, September 2002.

[19] S. Umeyama, “Least-Squares Estimation of Transformation Parameters Between Two Point Patterns,” IEEE. Trans. Pattern Analysis and Machine Intelligence, vol. 13, no. 4, pp. 376-380, April 1991.

[20] L. Torresani, D. B. Yang, E. J. Alexander, C. Bregler, “Tracking and Modeling Non-Rigid Objects with Rank Constraints,” Proc. of the IEEE Conf. Computer Vision and Pattern Recognition, vol. 1, pp. I-493 - I-500, December 2001.

[21] T. Morita and T. Kanade, “A Sequential Factorization Method for Recovering Shape and Motion From Image Streams,” IEEE. Trans. Pattern Analysis and Machine Intelligence, vol. 19, no. 8, pp. 858-867, August 1997.