臺灣博碩士論文加值系統

乳癌是台灣婦女第二高發生的癌症。根據統計，台灣地區每年約有5000名乳癌新診斷個案，其中有1400人死於乳癌，對國內婦女生命造成重大威脅。婦女除應定期自我檢查，當發覺有癌發的徵兆時，應先由專科醫師理學檢查，必要時安排進一步檢查，以決定病灶的嚴重度，然後選擇適當的治療方式。以期達到早期發現、早期治療的目的。
在乳房篩檢的過程中，病理組織切片檢查是乳癌診斷的黃金標準，除了能確定病理診斷外，另一重要目的為檢測乳癌預後因子，本研究希望利用乳房組織顯微鏡切片影像，透過影像處理技術的應用，分析與乳癌發生之可能相關組織型態及各種與腫瘤相關之特徵參數，藉以提供臨床醫師在乳癌診斷時之輔助，並應用於乳癌篩檢上。
研究過程中使用乳房組織切片之CCD數位影像作分析，透過影像分割的技術，包括色彩模型的轉換、臨界值的決定、影像型態學處理、相鄰元素標記法等影像處理步驟，分割出乳房導管，並取得導管輪廓及乳房導管之內、外緣區域影像，透過乳房導管特徵參數的擷取，計算影像中乳房導管比例、最大導管面積、乳房導管厚度、乳房導管內、外緣區域個數以及乳房導管分割影像紋理參數，並將其導入貝式網路中，藉此分析判斷組織切片影像是否有癌化的情況，提供病理醫師乳癌組織之形態分析與初步診斷。

The breast cancer is the second leading cancer for Taiwan woman. According to the statistics from the ministry of health, nearly 5000 cases of breast cancer are diagnosed every year in Taiwan. Among them, 1400 peoples die as a result of the breast cancer. This message causes great threat to woman’s life in Taiwan. To identify the disease as early as possible, periodically checking herself by touching her breast is a necessary and preliminary procedure to identify the disease. She should be further examined by an experienced doctor once she discovers possible symptoms of breast cancer. The severity of the disease can thus be identified as early as possible. Doctor can determine suitable and efficient treatments. During the examination of the breast cancer, examination with histology slice is a gold standard for the diagnosis of the cancer. It not only can provide the confirmation to breast biopsy in pathology, but the prognosis factor to examine the cancer.
The major goal of this research is to apply a series of image operations on the breast-biopsy-slide images fetched from microscopes, and analyze possible histology types and feature parameters related to the tumor. Therefore, it can help clinicians diagnose the disease, and apply the system to examine breast cancer, even reduce their load.
In this study, histology-slide images of breast fetched from CCD camera were used for experimental samples so that region of interest (i.e. duct) can be isolated form background. A series of image operations including color space transformation, thresholding, mathematical morphology, and connected component labeling were performed to extract the duct of breast from the histology-slide images. Next, various shape features including the area rate of the duct, the maximum area of the duct, the thickness of the duct, the number of inner and outer contours were extracted. Texture parameters such as entropy, contrast, and energy were also measured in order to analyze the complexity of the texture. The above features and parameters wee input into Bayes network where supervised learning were conducted to classify whether the examined patient infects the breast cancer. Experimental results reveal that the proposed system can obtain better performance for various testing samples.

中文摘要 i
Abstract ii
誌謝 iv
目錄 v
圖表目錄 ix
表格目錄 xi
第一章 導論 1
1.1研究動機 1
1.2研究背景及相關研究 5
1.2.1國內外相關研究現況 7
1.2.1-1 國外相關研究現況 7
1.2.1-2 國內相關研究現況 8
1.2.1-2參考文獻之評述 9
1.3研究目的 9
1.4論文架構 11
第二章 基礎理論 12
2.1正常的組織學 12
2.2乳癌的臨床表徵 13
2.3常見的良性乳房疾病 13
2.3.1乳房發炎及乳房膿瘍 14
2.3.2纖維囊腫病變 14
2.4乳房癌瘤 15
2.4.1管道癌瘤 15
2.4.2小葉性癌瘤 16
2.4.3最常見之乳癌 16
2.5乳癌之臨床分期 17
2.6病理組織學切片 20
2.7病理組織學切片成像 21
2.8色彩空間 22
2.8.1 RGB彩色模型 22
2.8.2 灰階轉換 23
2.9影像處理技術 24
2.9.1影像分割 24
2.9.1-1臨界值法 24
2.9.1-2形態學的修正 25
2.9.2 紋理分析 26
2.10貝氏網路 27
2.11評估方式 28
第三章 研究架構及方法 30
3.1研究影像與設備 30
3.1.1研究影像 30
3.1.2研究設備 31
3.2研究流程 31
3.3研究步驟 33
3.3.1影像分析及前處理 33
3.3.2乳房導管分割 36
3.3.3導管特徵參數及紋理分析 47
3.4系統診斷流程 53
3.5系統評估方式 54
第四章 實驗結果與討論 56
4.1乳房組織切片影像分割結果 56
4.1.1灰階轉換 56
4.1.2去除背景之處理 57
4.1.3去除紅血球之處理 58
4.1.4去除微小細胞核之處理 59
4.1.5乳導管內、外緣之擷取 61
4.2貝式網路訓練和測試 62
4.3系統介面 65
第五章 結論與未來展望 66
5.1結論 66
5.2未來展望 67
參考文獻 68

[1] N. H. ANDERSON, P. W. HAMILTON, P. H. BARTELS,D. THOMPSON, R. MONTIRONI and J. M. SLOAN,”Computerized scene segmentation for the discrimination of architectural features in ductal proliferative lesions of the breast, ” Journal of pathology, vol. 181: 374–380 (1997)

[2] P. Bamford and B. Lovell, “Unsupervised cell nucleus segmentation with active contours,” Signal Processing, 203-213,1998

[3] M. J. Bottema and J. P. Slavotinek,” Detection and classification of lobular and DCIS (small cell) microcalcifications in digital mammograms,” Pattern Recognition Letters vol: 21, Issue: 13-14, pp. 1209-1214, December, 2000

[4] H. C. Chen, W. J. Chien, and S. J. Wang, “Contrast-Based Color Image Segmentation,” IEEE vol. 11, Issue 7, Page(s):641 – 644, July 2004

[5] G. R. C and R. E. Woods, “Digital Image Processing / 2nd edition,” Prentice Hall, 2001.

[6] M. E. Celebi and Y. Alp Aslandogan Dept. of Computer Science and Engineering, University of Texas at Arlington Arlington, TX 76019-0015 U.S.A，Paul R. Bergstresser Department of Dermatology,U. of Texas Southwestern Medical Center Dallas, TX 75390-9003 U.S.A，”Mining Biomedical Images with Density-based Clustering”

[7] C. Demir and B. Yener,”Automated cancer diagnosis based on histopathological images: a systematic survey,”technical report, rensselaer polytechnic institute, department of computer science, TR-05-09.

[8] A. N. Esgiar, R. N. G. Naguib and S. Member,” Microscopic Image Analysis for Quantitative Measurement and Feature Identification of Normal and Cancerous Colonic Mucosa,”IEEE Trans. on information technology in biomedicine, vol. 2, no. 3, September 1998

[9] N. Esgiar, A. Naguib, R Sharif, B Bennett and M. Murray,” Microscopic image analysis for quantitative measurement and feature identification of normal and cancerous colonic mucosa,”Information Technology in Biomedicine, IEEE Trans. on vol 2, Issue 3, Page(s):197 – 203, Sept. 1998
[10] G. Elisabet; L. Xavier; S. Maria and M. Joan,” Computer aided diagnosis with case-based reasoning and genetic algorithms,”Knowledge-Based Systems vol: 15, Issue: 1-2,pp. 45-52, January, 2002

[11] H. Gao, W. C. Siu, and C. H. Hou,“Improved Techniques for Automatic Image Segmentation,” IEEE Trans. on vol. 11, Issue 12, Page(s):1273 – 1280,Dce.2001

[12] H. Gao, W. C. Siu and C. H Hou,”Improved Techniques for Automatic Image Segmentation,”IEEE Trans. On Circuit and Systems for Video Technology.

[13] J. GIL,* H. WU,B. Y. WANG,” Image Analysis and Morphometry in the Diagnosis of Breast Cancer ,”microscopy research and technique 59:109–118 (2002)

[14] C. Katsinis, S. Petushi, C. Coward, A. Tozeren and F. Garcia, "Automated Identification of Microstructures on Histology Slides,"IEEE International Symposium on Biomedical Imaging, Arlington, VA, pp. 424-427, April 15-18, 2004


[15] K. J. Khouzani and H. S. Xadeh, “Multiwavlet grading of pathological imagesof prostate,” IEEE Trans. On Biomed. Engineering, vol. 50, no.6, pp.697-704, June 2003.

[16] C. Katsinis, S. Petushi, C. Coward, A. Tozeren and F. Garcia, "Automated dentification of Microstructures on Histology Slides," IEEE International Symposium on Biomedical Imaging, Arlington, VA,, pp. 424-427, April 15-18, 2004

[17] R. Nock and F. Nielsen,“Statistical Region Merging,” IEEE Trans. on pattern analysis and machine intelligence, vol. 26, no. 11, November 2004

[18] C, Pan, C. X. Zheng and H.-J. Wang, “Robust color image segmentation based on mean shift and marker-controlled watershed algorithm,” Machine Learning and Cybernetics, 2003 International Conference on , Volume: 5 , Pages:2752 - 2756 Vol.5, 2-5 Nov. 2003

[19] S. Petushi,C. Katsinis,C. Coward,F. Garcia,A. Tozeren,”Automated identification of microstructures on histology slides,”Biomedical Imaging: Macro to Nano, 2004. IEEE International Symposium on 15-18 April 2004 Page(s):424 - 427 Vol. 1

[20] C. Ortiz de Solorzano, R. Malladi, SA Lelièvre and SJ Lockett. “Segmentation of nuclei and cells using membrane related protein markers,” Journal of Microscopy 201 (3), March 2001, 404–415

[21] F. Y. Shih * and S. Cheng ,“Adaptive mathematical morphology for edge linking,” Inf. Sci. 167(1-4): 9-21 (2004)

[22] B. Sahiner, H. P. Chan; Petrick, N., D. Wei, M. A. Helvie,D. D. Adler and M. M. Goodsitt,”Classification of mass and normal breast tissue: a convolution neural network classifier with spatial domain and texture images,”Medical Imaging, IEEE Trans, on vol 15, Issue 5, Page(s):598 – 610, Oct. 1996

[23] P. Viehweg,A. Heinig,D. Lampe,J. Buchmann and S. H. Heywang-Köbrunner,” Retrospective analysis for evaluation of the value of contrast-enhanced MRI in patients treated with breast conservative therapy,”Magnetic Resonance Materials in Biology, Physics and Medicine Vol: 7, Issue: 3, December 14, 1998, pp. 141-152

[24] B. Weyn, G. Wouwer,A. Daele,P. Scheunders,D. Dyck,E. Marck and W. Jacob1,” Automated Breast Tumor Diagnosis and Grading Based on Wavelet Chromatin Texture Description ,” Cytometry vol 33, Issue 1 , Pages 32 - 40

[25] H. S. Wu, J. Barba,* and J. Gil,”A Parametric Fitting Algorithm for Segmentation of Cell Images,”IEEE Trnas. on biomedical engineering, vol. 45, no. 3, March 1998

[26] 于南書，”最佳特徵選擇：乳房X光片腫瘤偵測”，成功大學資訊工程研究所碩士論文，2004

[27] 王惠暢，曾令民，吳秋文，雷永耀，彭芳谷，”乳房疾病新論”，2000.

[28] 巫宗昇，”利用彩色資訊之肝切片電腦影像分割”，成功大學資訊工程研究所碩士論文，1992。

[29] 季瑋珠，張金堅，”本土醫學資料庫之建立及衛生政策上之應用”，1993

[30] 國家衛生研究院 癌症研究組 TCOG乳癌研究委員會，”乳癌診斷與治療共識”，2004年10月

[31] 馮齡儀，”電腦輔助子宮頸抹片異常細胞辨識之初期研究”，中大學醫學工程研究所碩士論文，2004。

[32] 劉于綸，”細胞顯微影像分割與運動分析”，中央大學機械工程研究所碩士論文，2004

[33] 鐘兆春，方勝雄，”臨床生理學與病理切片技術”，1999

[34] 行政院衛生署，衛生統計資訊網之民國93年台灣地區主要死亡原因統計比較，http://www.doh.gov.tw/statistic/data/死因摘要/93年/表1.xls