由於超音波具有即時、低成本、容易使用以及非侵入性的優點，因此，在臨床上評估子宮內胎兒的生長情形大部分都使用超音波來進行檢查。透過量測胎兒之顱面結構，醫師可檢查出胎兒許多臨床上的症候群，如先天性畸形和唐氏症等。然而，超音波影像常因為訊號雜訊比太低、解剖構造不明顯等缺陷，容易造成醫師量測參數時，產生人為誤差及相當大的變異，此外，使得自動化胎兒顱面結構量測系統的開發變得相當困難。本論文為發展一套自動化三維超音波影像胎兒顱面結構量測系統，提出一個以模型為基礎之分割方法，完成自動化胎兒顱面結構分割，輔助醫師們量測八項具臨床意義之顱面結構參數。
在本論文所提出之分割方法，我們先藉由胎兒假體建立一個包含胎兒顱面的幾何形狀資訊與專家定義的標記點之樣板模型，將此樣板模型利用一模型形變分割架構，嵌合至三維超音波影像當中。在分割的過程中，我們提出一個以特徵為基礎之對位方法，成功達到自動化樣板模型位置初始化，並設計一個結合模型幾何形狀限制與影像特性之能量函數，進行三維樣板模型的形變，克服了上述複雜之超音波影像缺陷並完成胎兒顱面結構分割。我們將自動化量測之結果與專家手動量測之結果進行比較，從實驗當中可得到令人滿意的結果(影像分割平均誤差小於1.9個像素)。

In evaluation of the fetal growth in uterus, ultrasound is the most convenient and powerful tool in clinic due to its real time, low cost, easy to use and non-invasive nature. The doctors diagnose many anomaly syndromes of fetuses according to the measurement of fetal craniofacial structure. Since the ultrasound images have the disadvantages of low signal-to-noise ratio (SNR) and ambiguous anatomical structure, the human error and high variance may happen when the doctors measure the landmarks. However, it is also difficult to develop the automatic measurement system for fetal craniofacial structure. In this thesis, we propose a model-based segmentation approach to achieve automatic fetal craniofacial structure segmentation from 3D ultrasound images to assist the physician in measuring the eight meaningful craniofacial structure landmarks in clinics.
In the proposed segmentation method, we construct a template model which contains the geometric shape information of fetal craniofacial structure and ask an expert to define landmarks from a fetal phantom. Then, the template model is modified to the 3D ultrasound images by using a framework of model deformation. In the segmentation process, we propose a feature-based registration method to initialize template model position successfully and we design an energy function which combines the shape constraint of the template model and image properties to achieve 3D deformation of template model. We overcome the above-mentioned drawbacks of ultrasound images and successfully segment the fetal craniofacial structure. In the experiments, we achieve good consistency by comparing the automatic measurements with the manual ones.

[1] Ilan E. Timor-Tritsch, Ana Monteagudo, and Harris L. Cohen, Ultrasonography of the Prenatal and Neonatal Brain, second edition, McGraw-Hill Professional, 2001.
[2] Michael R. Kaus, Vladimir Pekar, Christian Lorenz, Roel Truyen, Steven Lobregt, and Jürgen Weese, “Automated 3-D PDM construction from segmented images using deformable models,” IEEE Transactions on Medical Imaging, Vol. 22, No. 8, pp. 1005 - 1013, August 2003.
[3] Olivier Ecabert, Jochen Peters, Hauke Schramm, Cristian Lorenz, Jens von Berg, Matthew J. Walker, Mani Vembar, Mark E. Olszewski, Krishna Subramanyan, Guy Lavi, and Jürgen Weese, “Automatic model-based segmentation of the heart in CT images,” IEEE Transactions on Medical Imaging, Vol. 27, No. 9, pp. 1189 - 1201, September 2008.
[4] Dinggang Shen, Yiqiang Zhan, and Christos Davatzikos, “Segmentation of prostate boundaries from ultrasound images using statistical shape model,” IEEE Transactions on Medical Imaging, Vol. 22, No. 4, pp. 539 - 551, April 2003.
[5] Wei Lu, Jinglu Tan, and Randall Floyd, “Automated fetal head detection and measurement in ultrasound images by iterative randomized Hough transform,” Ultrasound in Med. & Biol., Vol. 31, No. 7, pp. 929 - 936, 2005.
[6] Jinhua Yu,Yuanyuan Wang, Ping Chen, and Yuzhong Shen, “Fetal abdominal contour extraction and measurement in ultrasound images,” Ultrasound in Med. & Biol., Vol. 34, No. 2, pp. 169 - 182, 2008.
[7] http://www.gehealthcare.com/
[8] http://www.cirsinc.com/065_20_ultra.html
[9] Qingling Sun, John A. Hossack, Jinshan Tang, and Scott T. Acton, ”Speckle reducing anisotropic diffusion for 3D ultrasound images,” Computerized Medical Imaging and Graphics 28 (2004), pp. 461 - 470.
[10] William E. Lorensen and Harvey E. Cline, “Marching cubes: a high resolution 3D surface construction algorithm,” Computer Graphics, Vol. 21, No. 4, pp. 163 - 169, July 1987.
[11] Michael Garland and Paul S. Heckbert, “Surface simplification using quadric error metrics,” SIGGRAPH, pp. 209 - 216, 1997.
[12] http://meshlab.sourceforge.net/
[13] Chenyang Xu and Jerry L. Prince, “Snakes, shapes, and gradient vector flow,” IEEE Transactions on Image Processing, Vol. 7, No. 3, pp. 359 - 369, March 1998.
[14] Maurizio Pilu, Andrew W. Fitzgibbon, and Robert B. Fisher, “Ellipse-specific direct least-square fitting,” IEEE International Conference on Image Processing, Lausanne, pp. 599 - 602, September 1996.
[15] B.S. Manjunath and W.Y. Ma, “Texture features for browsing and retrieval of image data,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 18, No. 8, pp. 837 - 842, August 1996.
[16] Robert W. Sumner and Jovan Popovi´c, “Deformation transfer for triangle meshes,” SIGGRAPH, pp. 399 - 405, 2004.
[17] William H. Press , Saul A. Teukolsky , William T. Vetterling , and Brian P. Flannery, Numerical Recipes in C, second edition, The Press Syndicate of the University of Cambridge, 1992.