幼兒先天心臟血管疾病近年來一直是我國幼兒早亡的主要原因，發展良好的診斷工具，一直是研究幼兒先天心血管疾病重要課題；血管內超音波(Intravascular Ultrasound , IVUS)能即時獲得血管管腔及血管壁形態的高解析度影像，相當適合用來做血管的動態功能分析，然而血管內超音波只提供血管橫斷截面的資訊，卻無法知悉在人體內解剖的相對位置， 本研究將結合X光血管攝影術(X-ray angiography)以提供血管的空間定位，來建立一套血管阻塞定位系統，讓臨床醫師在獲得高解析度IVUS影像的同時，也能清楚了解相對應的解剖位置，作為診斷時的參考依據。另外本研究也利用壓克力來製作血管假體，之後經由模擬實驗來驗證此系統的準確性及穩定性。

　　本研究共分析了五位平均年齡10歲男童的資料，包含了20個區段的IVUS影像，比較電腦自動搜索的血管管腔邊界和專家手繪的結果，平均相關度為0.98顯示利用主動輪廓模型來自動搜索，確實能準確的搜索到內膜管腔邊界進而取代手繪邊界的描繪方法。

　　研究中使用壓克力材料製作血管假體，來評估定位系統之準確性。所重建出的血管假體三維模型與實際血管假體型態是符合的，血管假體在回拉定位的過程中，距離解析度可調範圍為1 mm/frame～50μm/frame，實驗中以1 mm/frame 距離解析度下，即可準確定位到假體位置。而在面積量測上平均誤差率為0.53 mm2 ，其誤差發生的原因主要是在血管假體缺口處，此處為一矩形的缺口，直角的部份在影像上會產生不連續點，因此當做邊界搜尋時會產生誤差。

　　Infant natural coronary artery disease is the main cause for infant early mortality in recent years in our country. Thus, the development of a diagnostic tool is an important issue for studying infant natural coronary artery disease. Intravascular Ultrasound(IVUS) is able to generates images immediately with a high temporal and spatial resolution of vessel wall morphology, so that it’s suitable for dynamic study of functional properties of vessels. However, IVUS only provides the information of vessel cross section. It’s unable to get the location of dissect inside the human bodies.This research combines X-ray angiography with IVUS to develop a vascular lesion localization system for clinical uses. Thus, the clinical physicians can get the images with high resolution and the dissect location simultaneously. The artificial vessels phantom are made of acrylic were used to check the validity of above system.

　　This research analyzes five subjects (boys, mean age 10 years old), including 20 IVUS sequential images; comparing the results of computer-detection and observer-defined for lumen boundary correlated very well (r=0.98), which means it can accurately search automatically for lumen boundary instead of manually tracing by experts.

　　The artificial vessel phantom made of acrylic evaluate the accuracy of above system. The artificial vessel phantom 3-D model could be fitted with the actual vessel phantom. On the process of artificial vessel phantom pullback position, the distance resolution range is 1mm / frame ～ 50μm/frame. In the experiment, it can position the artificial vessel precisely with 1mm / frame distance resolution. The average area error rate is 0.53mm2. The main cause for the error is artificial vessel breach. It’s a rectangular breach; the right angle part causes uncontinuous points on images. Therefore, the error happened in boundary detection.

[1]. 中華民國心臟病兒童基金會 http://www.ccft.org.tw

[2]. S. Nissen, J. Gurley, D. Booth, et al. “Spectrum of intravascular ultrasound findings in atherosclerosis: wall morphology and lumen shape in CAD patients,” J. Am Coll Cardiol, vol.11, 1991.

[3]. 林茂村, 人體生理學, 文京圖書有限公司, 1998

[4]. 莊克士, 醫學影像物理學, 合記圖書出版社, 1998.

[5]. N. Benjanmin, L. Antonio, M. James et. Al. “
Coronary artery imaging with intravascular high frequency ultrasound,” Circulation, pp. 1575-1585,1990.

[6]. F. W. Kermkau. Diagnostic Ultrasound, 4th ed, W. B Saunders Company, 1993.

[7]. J. W. Hole, and K. A. Koos, Human Anatomy, published by Wm. C. Brown Publishers,1991.

[8]. Z. Xiangmin, C. McKay, and M. Sonka, “Tissue characterizationin intravascular ultrasound images, ” IEEE Transactions on Medical Imaging, vol. 17, no.6 , pp.889 –899, 1998.

[9]. A. Oshima, M. Ochiai, and S. Takeshita, “Serial automated three-dimensional intravascular ultrasound analysis of the self-expanding Radius stent,” American Journal of Cardiology, vol. 85, no.3, pp. 388-91, 2000.

[10]. J. Hu, and X. Hu “Application of median filter to speckle suppression in intravascular ultrasound images,” Intelligent Information System, pp.302-306, 1994.

[11]. P. A. Brathwaite, K. B. Chandran, D. D. McPherson, E. L. Dove, “Lumen detection in human IVUS images using region-growing,” Computers in Cardiology, pp. 37 –40, 1996.

[12]. A. Mojsilovic, M. P. N. Popovic, N. Amodaj, R. Babic, and M. Ostojic, “Automatic segmentation of intravascular ultrasound images：A texture-based approach,” Annals of biomedical engineering, vol. 25, pp. 1059-1071, 1997.

[13]. C. V. Birgelen, G. S. Mintz, D. P. Foley, V. D. Giessen, S. G. Airiian, T. C. Roelandt, P. J. Feyter, and P. W. Serruys, “Electrocardigram-gated intravascular ultrasound image acquisition coronary stent deployment facilitates on-line three-dimensional reconstruction and automated lumen quantification,” JACC, vol.30,no.2, pp.436-443,1997.

[14]. A.Wahle, P. M. Prause, C. DeJong, and M. Sonka, ”Geometrically Correct 3-D Reconstruction of Intravacular Ultrasound Image by Fusion with Biplane Angiography — Methods and Validation,” IEEE Transaction on Medical Imaging , vol. 18, no. 8, p.p. 256-274, 1999.

[15]. M. J. Vonesh, C. Kequing, and M. Radvany, “Digital subtraction for noise reduction in intravascular ultrasound data,” Computers in Cardiology, pp. 329-332, 1990.

[16]. R. C. Gonzalez and Richard E. Woods. Digital Image Processing. Addison-Wesley Publishing Company, 1992.

[17]. A. Kass, M. Witkin, and D. Terzopoulos, ” Snakes:Active
Countour Models,” Int. J. Comput. Vision, pp321-331, 1998.

[18]. M. Sonka, et al, Image Processing -Analysis and Machine Vision, 滄海書局, pp. 374-380, 1999.

[19]. M. Galloway, “Texture analysis and classification using gray length run lengths,” Comput. Graphics Image Process. 4, pp.172-199, 1974.

[20]. L. D. Cohen, “On Active Contour Models and
Balloons,” CVGIP: Image Understand., vol. 53, pp. 211-218, Mar. 1991.

[21]. 繆紹綱, 數位影像處理~活用Matlab, 全華科技圖書股份有限公司印行, 台北市, 1999.

[22]. 吳祈德, “血管內超音波影像邊界偵測及三度空間血管模型之建立”，成功大學醫學工程研究所碩士論文,民國九十年六月.

[23]. 林子敬, “利用血管內超音波和雙平面血管X光攝影建立三度空間血管模型”，成功大學醫學工程研究所碩士論文,民國九十二年
六月.