在乳房攝影上，大量的形態特徵及模糊邊緣會導致在正常區域內區分出腫塊是一項艱鉅的任務，且在目前的乳癌篩檢中，放射科醫師會誤判的機率為10% - 30%。基於此項原因，目前已存在許多的電腦輔助檢測系統，藉此幫助醫師檢測出可能導致乳癌的病灶。這些電腦輔助檢測系統僅能做為參考，最後結果還是由醫師作出決定。近期的研究也著重於將電腦輔助檢測系統當作一種工具使用，以提高乳房攝影檢測乳癌的準確性。因此在此研究中，我們提出一種新的特徵擷取方法用來增進乳房攝影中檢測腫塊的效能。

在腫瘤偵測系統中，可分成五個步驟。首先，在我們的研究中，僅處理乳房的
部分，因此我們會先去除胸大肌。由於乳房攝影上，實質紋理與胸大肌相似，所以先去除胸大肌可減低系統的假陽性率。接著，我們利用形態濾波器來抑制血管及結構雜訊，並採用模板比對方法尋找可疑的區域。根據可疑區域的大小，定義出適應性的矩形感興趣區域(ROI)。我們採用光密度共生矩陣及光度特徵來定義ROI的特徵，其中光密度共生矩陣為加入背景資訊的灰階共生矩陣。另外，也定義了一組新的特徵擷取方式來描述ROI 上的特徵。最後，我們利用基於逐步方法之線性區別函數得到區別函數並利用此函數來區分腫瘤與正常組織。

我們從數位乳房攝影資料庫中取出358個案例來進行實驗，其中包含了180個惡
性腫瘤、128個良性腫瘤及50個正常案例。我們提出的方法的平均靈敏度為97.3%，假陽性率為8.8%及ROC曲線下的面積為0.976。尤其，在乳房緻密度為第三級時，敏感度為97.6%且假陽性率為10.2%。而在乳房緻密度為第四級時，敏感度達100%且假陽性率為8.9%。結果證明，我們達到了令人滿意的檢測性能且在靈敏度與假陽性率之間取得平衡。

Distinguishing masses from normal regions is a di cult task owing to the great numbers of morphological characteristics and ambiguous margins which are also present in images. It has been shown that 10% - 30% of the tumors are missed by the radiologists in current breast cancer screenings. Due to this reason, Computer-Aided Detection (CAD) systems have been developed to aid radiologists in detecting mammographic lesions that may indicate the presence of breast cancer. These systems act only as a second reader and the final decision is made by the radiologists. Recent studies have also shown that CAD detection systems as an aid have improved radiologists' accuracy of detection of breast cancer. In this thesis, two new feature extraction methods are provided to improve the performance of mass detection in CAD system.

The mass detection method is proposed which consist of five steps. First, remove the pectoral muscle because the breast part should be processed in our research. Removing the pectoral muscle could avoid false positives when detecting masses due to the similarity of the texture of mammographic parenchyma and the pectoral muscle. Second, a morphological filter is used to suppress both blood vessels and structural noises. Afterward, a template matching method is used to find the suspicious area. An adaptive square region of interesting (ROI) are determined according to the suspicious area's size. Total 69 descriptors including optical density co-occurrence matrix and photometric feature are introduced to de ne the characteristics of a ROI. The optical density co-occurrence matrix which refers to the background information refines Haralicks' texture feature. In addition, there is other new combination of features which also describes the characteristics of a ROI in our system. Finally, the linear discriminant analysis (LDA) based on stepwise method classifier is utilized to determine the discriminant function which is used to distinguish masses and normal regions.

In our experiment, 358 cases containing 180 malignant tumors, 128 benign tumors and 50 normal cases from the Digital Database for Screening Mammography is conducted. The average sensitivity of the proposed scheme is 97.3%, false positive rate of 8.8% and the Az is 0.976 0.004. In particular, the sensitivity is 97.6% at false positive rate 10.2% in mammographic density 3 and the sensitivity is 100% at false positive rate 8.9% in mammographic density 4. The results proved that the proposed detection scheme achieves satisfactory detection performance and preferable compromises between sensitivity and false positive rates.

[1] " Image Retrieval in Medical Applications." [Online]. Available: http:
//irma-project.org.
[2] " NCI Cancer Fact Sheets. (2012)." [Online]. Available: http://www.cancer.gov/
cancertopics/types/breast
[3] " WHO Cancer Fact Sheets. (2012)." [Online]. Available: http://www.who.int/
mediacentre/factsheets/fs297/en/index.html
[4] J. Ball and L. Bruce, " Digital mammogram spiculated mass detection and spicule
segmentation using level sets," in Engineering in Medicine and Biology Society,
2007. EMBS 2007. 29th Annual International Conference of the IEEE, August 2007, pp. 4979 -4984.
[5] I. Bankman, Ed., Handbook of Medical Image Processing and Analysis, 2008, vol. 2.
[6] R. Bellotti, F. De Carlo, S. Tangaro, G. Gargano, G. Maggipinto, M. Castel-
lano, R. Massafra, D. Cascio, F. Fauci, G. Magro, R .and Raso, G. Lauria, A
.and Forni, S. Bagnasco, P. Cerello, E. Zanon, S. Cheran, E. Lopez Torres, U. Bot-
tigli, G. Masala, P. Oliva, M. Retico, A .and Fantacci, R. Cataldo, I. De Mitri,
and G. De Nunzio, " A completely automated cad system for mass detection in a
large mammographic database." Med Phys, vol. 33, no. 8, pp. 3066-3075, August 2006.
[7] C. Berman, " Recent advances in breast-specific imaging," Cancer Control, vol. 14, no. 4, pp. 338-349, October 2007.
[8] R. E. Bird, T. W. Wallace, and B. C. Yankaskas, "Analysis of cancers missed at
screening mammography," Radiology, vol. 184, pp. 613-617, September 1992.
[9] J. Bozek, M. Mustra, K. Delac, and M. Grgic, "A survey of image processing
algorithms in digital mammography," Mult. Sig. Process. and Commun, vol. 231, pp. 631-657, 2009.
[10] R. F. Brem, J. W. Ho_meister, G. Zisman, M. P. DeSimio, and S. K. RogersSeptember, " A computer-aided detection system for the evaluation of breast cancer by mammographic appearance and lesion size," Amer. J. Roentgenol, vol.
184, no. 3, pp. 893-896, March 2005.
[11] R. Brem, J. Rapelyea, G.Zisman, J. Ho_meister, and M. Desimio., "Evaluation
of breast cancer with a computer-aided detection system by mammographic appearance and histopathology," Cancer, vol. 104, no. 5, pp. 931-935, September
2005.
[12] H. Chan, D. Wei, M. Helvie, B. Sahiner, D. Adler, M. Goodsitt, and N. Petrick,
"Computer-aided classi_cation of mammographic masses and normal tissue: Linear discriminant analysis in texture feature space," Phys Med Biol., vol. 40, no. 5,
pp. 857-876, May 1995.
[13] H. Cheng, X. Shi, R. Min, L. Hu, X. Cai, and H. Du, "Approaches for automated detection and classification of masses in mammograms," Pattern Recognition, vol. 39, no. 4, pp. 646-668, June 2006.
[14] S. Ciatto, N. Houssami, D. Gur, R. Nishikawa, R. Schmidt, C. Metz, J. Ruiz,
S. Feig, R. Birdwell, M. Linver, J. Fenton, W. Barlow, and J. Elmore, "Computer-
aided screening mammography," New England J. Med, vol. 357, no. 1, pp. 83-85,
July 2007.
[15] A. Dominguez and A. Nandi, " Enhanced multi-level thresholding segmentation and rank based region selection for detection of masses in mammograms," in Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International
Conference on, vol. 1, April 2007, pp. I-449 -I-452.
[16] N. Eltonsy, G. Tourassi, and A. Elmaghraby, "A concentric morphology model for the detection of masses in mammography," Medical Imaging, IEEE Transactions on, vol. 26, no. 6, pp. 880 -889, June 2007.
[17] M. Eltoukhy, I. Faye, and B. Samir, "Automatic detection of breast masses in digital mammograms using pattern matching," in Biomedical Engineering and
Sciences (IECBES), 2010 IEEE EMBS Conference on, December 2010, pp. 73-76.
[18] X. Gao, Y. Wang, X. Li, and D. Tao, "On combining morphological component
Analysis and concentric morphology model for mammographic mass detection,"
Information Technology in Biomedicine, IEEE Transactions on, vol. 14, no. 2, pp.
266 -273, March 2010.
[19] M. L. Giger, "Computer-aided diagnosis in radiology," Acad.Radiol, vol. 9, pp. 1-3, July 2002.
[20] R. C. Gonzalez and R. E. Woods, Digital Image Processing, 3rd ed. Prentice Hall, 2008.
[21] R. Haralick, "Statistical and structural approaches to texture," Proceedings of the IEEE, vol. 67, no. 5, pp. 786 - 804, May 1979.
[22] M. Heath, K. Bowyer, D. Kopans, R. Moore, and P. K. Jr., "The digital database for screening mammography," 2000.
[23] S. Javadi and K. Faez, "Finding suspicious masses of breast cancer in mammography images using particle swarm algorithm and its classification using fuzzy methods," in Computer Communication and Informatics (ICCCI), 2012 International Conference on, January 2012, pp. 1 -5.
[24] K. Kerlikowske, P. A. Carney, and B. Geller, "Performance of screening mammography among women with and without a first-degree relative with breast cancer, "Annals Int. Med., vol. 133, no. 11, pp. 855-863, December 2000.
[25] T. M. Kolb, J. Lichy, and J. H. Newhouse, "Comparison of the performance of screening mammography, physical examination, and breast us and evaluation of factors that inuence them: an analysis of 27,825 patient evaluations." Radiology, vol. 225, no. 1, pp. 165-175, October 2002.
[26] G. Kom, A. Tiedeu, and M. Kom, " Automated detection of masses in mammograms by local adaptive thresholding," Computers in Biology and Medicine, vol. 37, no. 1, pp. 37-48, January 2007.
[27] D. B. Kopans, "The positive predictive value of mammography," AJR. Am. J. Roentgenol, vol. 158, pp. 521-526, October 1992.
[28] C. H. Lee, "Screening mammography: proven benefit, continued controversy." Radiol. Clinics North Am, vol. 40, no. 3, pp. 395-407, May 2002.
[29] D. Michelle, M. Jaclyn, F. William, S. Susan, and A. Louise, "Global trends in breast cancer incidence and mortality 1973v1997," Int. J. Epidemiol, vol. 34, no. 2,
pp. 405-412, April 2005.
[30] A. C. of Radiology, ACR BI-RADS-Mammography, Ultrasound & Magnetic Resonance Imaging. 4th ed. American College of Radiology (Reston, VA, 2003).
[31] E. D. Pisano, C. Gatsonis, E. Hendrick, M. Ya_e, J. K. Baum, S. Acharyya,
E. F. Conant, L. L. Fajardo, L. Bassett, C. D'Orsi, R. Jong, and M. Rebner, "Diagnostic performance of digital versus _lm mammography for breast-cancer
screening," New England J. Med, vol. 353, no. 17, pp. 1773-1783, October 2005.
[32] G. Rabottino, A. Mencattini, M. Salmeri, F. Caselli, and R. Lojacon, "Mass contour extraction in mammographic images for breast cancer identification," in 16th IMEKO TC4 Symposium Exploring New Frontiers of Instrumentation and Methods for Electrical and Electronic Measurement, July 2008.
[33] R. M. Rangayyana, F. bio J. Ayresa, and J. L. Desautels, "A review of computer-
Aided-diagnosis of breast cancer: Toward the detection of subtle signs," Journal of the Franklin Institute, vol. 344, no. 3-4, pp. 312-348, July 2007.
[34] M. Sameti, R. Ward, J. Morgan-Parkes, and B. Palcic, "Image feature extraction in the last screening mammograms prior to detection of breast cancer," Selected Topics in Signal Processing, IEEE Journal of, vol. 3, no. 1, pp. 46-52, February 2009.
[35] M. P. Sampat, M. K. Markey, and A. C. Bovik, Computer-aided detection and diagnosis in mammography, in handbook of image and video processing ed., A. C. B. 2nd, Ed., New York: Academic, 2005.
[36] J. Tang, R. Rangayyan, J. Xu, I. El Naqa, and Y. Yang, "Computer-aided detection and diagnosis of breast cancer with mammography: Recent advances, " Information Technology in Biomedicine, IEEE Transactions on, vol. 13, no. 2, pp.236-251, March 2009.
[37] T. Terada, Y. Fukumizu, H. Yamauchi, H. Chou, and Y. Kurumi, "Detecting mass and its region in mammograms using mean shift segmentation and iris filter," in Communications and Information Technologies (ISCIT), 2010 International Symposium on, October 2010, pp. 1176 -1179.
[38] C. Varelaa, P. G. Tahocesb, A. J. M_endez, M. Soutoa, and J. J. Vidala, "Computerized detection of breast masses in digitized mammograms," Computers in Biology and Medicine, vol. 37, no. 2, pp. 214-226, February 2007.
[39] J. Wei, B. Sahiner, L. M. Hadjiiski, H. P. Chan, N. Petrick, M. A. Helvie, M. A. Roubidoux, J. Ge, and C. Zhou, "Computer-aided detection of breast masse on full field digital mammograms," Med. Phys, vol. 32, no. 9, pp. 2827-2837, September 2005.
[40] S. Xu and C. Pei, "Hierarchical matching for automatic detection of masses in
mammograms, " in Electrical and Control Engineering (ICECE), 2011 International Conference on, September 2011, pp. 4523 -4526.
[41] H. J. Yoon, B. Zheng, and D. P. Chakraborty, "Evaluating computer-aided detection algorithms," Med. Phys., vol. 34, no. 6, pp. 2024-2048, June