對胸部X光影像進行分析一直是研究醫學影像的一個主要項目。由氣管癌、支氣管癌、肺癌在2008年的世界十大死因排名第七即可知其在研究上的重要性。肺臟區域內肋骨交疊是醫師診斷肺癌時的主要干擾，82%~95%肺癌的誤診可歸因於鄰近的肋骨或鎖骨遮擋患部。於此我們提出一個肋骨切割演算法，藉由正確的標示肋骨位置，協助醫生在診斷時能將精力專注於對付病灶。
本論文中，我們提出一個非監督式的自動切割肋骨演算方法，此方法是基於事先預備好的肺域遮罩以進行肋骨切割。部分文獻面對同類型問題所採用的方法是基於數學模型擬合切割肋骨，我們提出另一種方式去獲取好的切割結果。我們並未採用現有數學模型作為模板，而是根據肋骨趨勢設計追蹤規則，從定位好的遮罩邊緣追蹤數條肋骨邊緣以形成模板，此方式所產生模板更貼近真實肋骨邊緣。因為僅僅使用單一模板無法有效描述所有肋骨，故採用複數模板進行比對，再藉由追蹤所有模板重疊位置結果，以保留邊緣特徵較強及統計特徵合適的肋骨邊緣，後續再將剩餘邊緣兩兩配對。最後，對每根配對完成肋骨分別執行主動輪廓模型演算法，可得到最後切割結果。上述演算法流程，在不同階段分別使用邊緣及灰階特徵，由起始粗略位置逐步進展至最後切割結果。整個演算法經由多張不同胸部X光影像驗證，可獲得相當不錯的切割結果，足以顯示所提方法具有很好的可靠度。

Analysis of the X-ray image is one of the major issues in medical image processing. Trachea, bronchus, lung cancers are the seventh leading cause of death worldwide in 2008. The superimpositions of normal anatomical structures often cause interference, and many, 82%~95%, of the missed lung cancers were partly obscured by overlying bones such as ribs or a clavicle. We propose a segmentation method for ribs in chest radiographs to assist doctors in diagnosis.
In this thesis, an unsupervised rib segmentation method is proposed based on existing mask of lung field. Some literatures deal with this problem based on a model fitting method. Instead of using a mathematical model as a template, we proposed another way achieving a good result. We utilize the designed rules of tracing from the border of mask of lung field to obtain a curve closer to the real edge. Because it is insufficient for using only one template to describe all ribs, we analyze all results of all templates in an image. After fine tuning, we keep routes with strong edge which satisfy all global features. The pairing procedure is performed on the edges. Finally, based on the paired edges as initial contour, active contour model is adopted to get the final result. The algorithm processes at different stages were characterized using edge and gray value, a coarse position is captured and gradually progressed to the final results. From the experimental results, the proposed algorithm can effectively segment the ribs in the chest radiographs.

[1] http://www.who.int/mediacentre/factsheets/fs310/en/
[2] Kenji Suzuki*, Hiroyuki Abe, Heber MacMahon, and Kunio Doi,“Image-Processing Technique for Suppressing Ribs in Chest Radiographs by Means of Massive Training Artificial Neural Network (MTANN)”Proceedings of the IEEE TRANSACTIONS on MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006
[3] Zhanjun Yue, Ardeshir Goshtasby, and Laurens V. Ackerman“Automatic Detection of Rib Borders in Chest Radiographs”Proceedings of the IEEE TRANSACTIONS on MEDICAL IMAGING, VOL. 14, NO. 3, SEPTEMBER 1995
[4] Marco Loog* and Bram van Ginneken,“Segmentation of the Posterior Ribs in Chest Radiographs Using Iterated Contextual Pixel Classification”Proceedings of the IEEE TRANSACTIONS on MEDICAL IMAGING, VOL. 25, NO. 5, MAY 2006
[5] Song Ya-Lin and Yang Yang,“Localization Algorithm and Implementation for Focal of Pulmonary Tuberculosis”Proceedings of the International Conference on Machine Vision and Human-machine Interface, April 2010
[6] John Canny,“A Computational Approach to Edge Detection”Proceedings of the IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. PAMI-8, NO6, NOVEMBER 1986
[7] Rafael C. Gonzalez and Richard E. Woods,“Digital Image Processing” Addison-Wesley Longman Publishing Co., Inc, Boston, Ma, 1992
[8] Michael Kass, Andrew Witkin, and Demetri Terzopoulos,“Snakes: Active Contour Models”Proceedings of the International Journal of Computer Vision, p. 321- 331, 1988
[9] Cheng-Hung Lai,“Unsupervised Segmentation of Lung Fields in Chest Radiographs by Multiresolution Fractal Feature Vector”, Master Thesis, National Tsing Hua University, Taiwan, 2010
[10] Wen-Li Lee, Yung-Chang Chen, and Kai-Sheng Hsieh,“Ultrasonic Liver Tissues Classification by Fractal Feature Vector based on M-band Wavelet Transform”Proceedings of the IEEE TRANSACTIONS on MEDICAL IMAGING, VOL22, NO. 3, MARCH 2003
[11] J. MACQUEEN,“Some Method for Classification and Analysis of Multivariate Observation”Proceedings of 5th Berkeley Symposium on Mathematical Statistics and Probability, University of California Press, p. 281–297, April 2009.
[12] Y. Boykov and G. Funka-Lea,“Graph cuts and efficient n-d image segmentation”Proceedings of the International Jornal of Computer Vision, 70(2), 109-131, 2006
[13] D.H. Ballard,“Generalizing the Hough Transform to Detect Arbitrary Shapes”Pattern Recognition, Vol.13, No.2, p.111-122, 1981
[14] Andrew B. Watson, “Spatial Standard Observer”, United States Patent Application Publication, US2006/0165311 A1
[15] Chenyang Xu and Jerry L. Prince, “Snakes, Shapes, and Gradient Vector Flow” Proceedings of the IEEE TRANSACTIONS on IMAGE PROCESSING, VOL7, NO. 3, MARCH 1998