ENHANCEMENT AND FEATURE SELECTION FOR CLASSIFICATION OF X-RAY MAMMOGRAM MICROCALCIFICATIONS
蔡明倫、傅家啟,李三剛
E-mail: [email protected]
ABSTRACT
MICROCALCIFICATIONS MAY APPEAR AS AN EARLY SIGN OF BREAST CANCER AND PLAY AN IMPORTANT ROLE IN DIAGNOSING MAMMOGRAM.IN GENERAL,THE MODELS OF DETECTING MICROCALCIFICATIONS HAVE DIFFERENT KINDS OF DETECTION SCHEMES AND VARIOUS FEATURE EXTRACTION METHODS.IN THIS THES -IS,A DIAGNOSTIC METHOD WITH TWO STAGE SCHEME AND DATA MINING BASED FEATURE EXTRACTION W -ITH HIGH PERFORMANCE CLASSIFIER IS DEVELOPED TO DETECT
MICROCALCIFICATIONS IN X-RAY MAM -MOGRAMS.THIS THESIS PRESENTS A COMPUTER-AIDED DIAGNOSIS (CAD) SYSTEM FOR THE AUTOMATIC DETECTION OF MICROCALCIFICATIONS IN DIGITIZED X-RAY MAMMOGRAMS.THE PROPOSED SYSTEM CONSI -STS OF TWO MAIN STEPS. FIRST,POTENTIAL MICROCALCIFICATION PIXELS ARE SEGMENTED BY A CLO -SED FORM CONSERVATIVE FILTER. THEN, INDIVIDUAL MICROCALCIFICATION IS TESTED BY THE FEAT -URES EXTRACTED FROM THE SPATIAL DOMAIN AND THE SPECTRAL DOMAIN. THE DISCRIMINATORY POWER OF THESE FEATURES IS ANALYZED VIA SEQUENTIAL FORWARD SELECTION METHOD. EXPERIMENT RESULTS SHOW THAT DATA MINING SCHEME IS SUPERIOR TO FEATURES WITHOUT BEING SELECTED. THE SUPPORT VECTOR MACHINE CLASSIFIER IS SUPERIOR TO THE GENERAL REGRESSION NEURAL NETWORK CLASSIFIER.
Keywords : MICROCALCIFICATIONS, FEATURE SELECTION, DATA MINING, CLASSIFIER Table of Contents
封面內頁 簽名頁 授權書一 iii 授權書二 iv 中文摘要 v 英文摘要 vi 誌謝 vii 目錄 viii 圖目錄 xi 表目錄 xiii 第一章 緒論 1 1.1研 究目的與動機 1 1.2研究範圍 2 1.3研究方法 2 第二章 文獻探討 3 2.1微鈣化檢測模式 3 2.1.1一階段影像檢測模式 3 2.1.2兩階 段影像檢測模式 4 2.2影像強化 6 2.2.1空間方法 6 2.2.2 統計方法 13 2.3特徵萃取 15 2.3.1空間域特徵 15 2.3.2頻率域特徵 20 2.4特徵組合資料探勘 21 逐次前饋式搜尋法(SFS) 21 2.5分類器 22 2.5.1 General Regression Neural Network (GRNN) 25 2.5.2 Support Vector Machine (SVM) 28 2.6績效衡量-ROC曲線 35 第三章 研究方法 39 3.1保守濾波器的設計 39 3.2特徵萃取的應 用 42 3.3資料探勘的概念 44 3.4 SVM分類器與GRNN分類器 44 第四章 實驗結果 46 4.1實驗設置及流程 46 4.1.1應用保守濾 波器 47 4.1.2擷取訓練資料集與測試資料集 48 4.1.3特徵萃取 49 4.1.4特徵選擇與績效評量 50 4.2實驗結果及分析 50 4.2.1保 守濾波器結果 50 4.2.2特徵選擇結果 52 第五章 結論與未來研究發展 62 5.1結論 62 5.2未來研究發展 64 參考文獻 66 REFERENCES
1. 吳智誠, "資料探勘於影像資訊之應用-以乳房微鈣化特徵處理為案例", 大葉大學工業工程學系碩士 班論文(指導教授:傅家啟，共同指導 教授:李三剛), 2001. 2. 繆紹綱, "數位影像處理-活用MATLAB", 全華科技圖書股份有限公司, 1999. 3. 羅強華, "類神經網路:MATLAB 的應 用", 清蔚科技股份有限公司, 2001. 4. BROWN, M., GRUNDY, W., LIN, D., CRISTIANINI, N., SUGENT, C., FUREY, T.,ARES, M., HAUSSL -ER, D., "KNOWLEDGE-BASED ANALYSIS OF MICROARRAY GENE EXPRESSION DATA USING SUPPORT VEC -TOR MACHINES", TECHNICAL REPORT, UNIVERSITY OF CALIFORNIA IN SANTA CRUZ, 1999. (SUBMIT -TED FOR
PUBLICATION). 5. HARALICK, R. M., SHANMUGAM, K., DINSTEIN, I., "TEXTURAL FEATURES FOR IMAGE CLASSIFICAT -ION", IEEE TRANS. SYSTEMS, MAN, AND CYBERNETICS,VOL. SMC-3, NO. 6, NOV. 1973. 6. HEYWANG-KOBRUNNER, S. H., SCHREER, I., DERSHAW, D. D., "DIAGNOSTIC BREAST IMAGING", THI -EME MEDICAL PUBLISHERS, 1997. 7. KIM, J. K., PARK, H. W., "STATISTICAL TEXTURAL FEATURES FOR DETECTION OF MICROCALCIFICA -TIONS IN DIGITIZED MAMMOGRAMS", IEEE TRANS. MED. IMAG.,VOL.18, NO. 3, PP. 231-238, MAR. 1999. 8. LECUN, Y., JACKEL, L. D., BOTTOU, L., BRUNOT, A., CORTES, C., DENKER, J. S., DRUCKER, H.,GUYON, I., MULLER, U. A., SACKINGER, E., SIMARD, P.,VAPNIK, V.,
"COMPARISON OF LEARN -ING ALGORITHM FOR HANDWRITTEN DIGIT RECOGNITION",IN FOGEL MAN-SOULIE, F. AND GALLINARI, P., EDITORS, PROCEEDINGS ICANN'95, VOLUME II, PAGES 53-60. EC2, 1995. 9. MORROW, W. M., PARANJAPE, R.
B., RANGAYYAN, R. M., DESAUTELS, J. E. L.,"REGION-BASED CO -NTRAST ENHANCEMENT OF MAMMOGRAMS", IEEE
TRANS. MED. IMAG., VOL. 11, NO. 3, PP. 392-406, SEPT. 1992. 10.NELLO, C. J., SHAVE T., "AN INTRODUCTION TO SUPPORT VECTOR MACHINES: AND OTHER KERNEL BASED LEARNING METHOD", CAMBRIDGE UNIVERSITY PRESS,2000. 11.OREN, M., PAPAGEORGIOU, C., SINHA, P., OSUNA, E., POGGIO, T.,"PEDESTRIAN DETECTION USING WAVELET TEMPLATES", THE PROCEEDINGS OF CVPR'97, JUNE 1997. 12.OSUNA, E., FREUND, R., GIROSI, F., "TRAINING SUPPORT VECTOR MACHINES:
AN APPLICATION TO FACE DETECTION", THE PROCEEDINGS OF CVPR'97, JUNE 1997. 13.PRINCIPE, J. C., EULIANO, N. R., LEFEBVRE, W. C., "NEURAL AND ADAPTIVE SYSTEMS: FUNDAME -NTALS THROUGH SIMULATIONS", JOHN WILEY AND SONS, 1999. 14.SHEN, L., RANGAYYAN, R. M., LEO DESAUTELS, J. E., "SHAPE ANALYSIS OF MAMMOGRAPHIC CALC
-IFICATIONS", IN PROC. FIFTH ANNU. IEEE SYMP.COMPUTER-BASED MEDICAL SYSTEMS, JUNE 1992, PP. 123-128. 15.SONKA M., HLAVAC, V., BOYLE, R., "IMAGE PROCESSING, ANALYSIS,AND MACHINE VISION",BROOKS /COLE PUBLISHING
COMPANY, 1999. 16.STRICKLAND, R.N. ,HAHN, H.I. , "WAVELET TRANSFORMS FOR DETECTION MICROCALCIFICATIONS IN MAMMOGRAMS", IEEE TRANS. ON MEDICAL IMAGING, VOL.15, NO.2,APRIL 1996. 17.VAPNIK, V., "THE NATURE OF STATISTICAL LEARNING THEORY", SPRINGER VERLAG, 1995. 18.VELDKAMP, W. J. H., KARSSEMEIJER, N.,
"NORMALIZATION OF LOCAL CONTRAST IN MAMMOGRAMS", IEEE TRANS. MED. IMAG., VOL. 19, NO. 7,PP. 731-738, JULY 2000. 19.YU, S., GUAN, L., "A CAD SYSTEM FOR THE AUTOMATIC DETECTION OF CLUSTERED MICROCALCIFIC -ATIONS IN DIGITIZED MAMMOGRAM FILMS", IEEE TRANS. MED. IMAG., VOL. 19, NO. 2, PP. 115 -126, FEB. 2000. 20.ZHENG, B., QIAN, W., CLARKE, L. P., "DIGITAL MAMMOGRAPHY MIXED FEATURE NEURAL NETWORK WITH SPECTRAL ENTROPY
DECISION FOR DETECTION OF MICROCALCIFICATIONS", IEEE TRANS. MED. IMAG., VOL. 15, NO. 5, PP. 589-597, OCT. 1996.
21 .HTTP://CHEMDIV-WWW.NRL.NAVY.MIL/6110/SENSORS/CHEMOMETRICS/GRNN. HTML 22.
HTTP://WWW.DOH.GOV.TW/STATISTIC/DATA/公佈欄資料檔/90 死因/90 年死因分析.DOC 23.
HTTP://WWW.DOH.GOV.TW/STATISTIC/DATA/公佈欄資料檔/90 死因/癌症8990.XLS
