( 您好!臺灣時間:2021/04/19 19:38
字體大小: 字級放大   字級縮小   預設字形  


研究生(外文):Yu-Ruei Huang
論文名稱(外文):The Study of Uterine Cervical Cellular Images Based on PC-based Cytopathologic Image Analysis System and Support Vector Machine
指導教授:陳 榮 靜
指導教授(外文):Rung-Ching Chen
外文關鍵詞:cytologyImage analysiscervical cancerSupport vector machine
  • 被引用被引用:2
  • 點閱點閱:369
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:51
  • 收藏至我的研究室書目清單書目收藏:0
最初我們利用顯微鏡及數位相機拍攝超柏氏抹片,共拍攝1814個細胞影像,每個細胞分別利用100、200與400倍的倍率拍攝,且依照四種診斷種類分類,分別為正常細胞,包括Superficial squamous cell (S Cell)、Intermediate squamous cell (I Cell)、parabasal squamous cell (P Cell)與異常細胞,包括Low grade squamous intraepithelial lesion (LSIL Cell)與high grade squamous intraepithelial lesion (HSIL)兩種類型癌症細胞。之後,經由病理學家討論一致通過之後,評選其中的503個放大倍率為400倍的細胞影像,供未來實驗。
我們設計的細胞影像分析系統,能自動批次輸入細胞影像,將彩色細胞影像轉成灰階影像,並針對13個主要的細胞型態參數,如細胞核大小、形狀、核質比與核染值做客觀的運算分析。我們利用支援向量機(support vector machine, SVM)對細胞影像分析系統所產生的數據做訓練(Training)與預測(predict)。
結果顯示我們評估後所選擇的細胞型態參數能正確的將異常細胞從正常細胞群中分辨出來(statistic significance p<0.001)。 SVM的預測結果能得到高準確率的分類(accuracy up to 94.43%, sensitivity: 100% and specificity: 98.5%, 沒有 false negative 的例子)。
Cytologic screening has been widely used for controlling the prevalence of cervical cancer. However, errors from sampling, screening and interpretation, still concealed some unpleasant results. This study aims to design a cellular image analysis system with feasible and available software and hardware for a routine cytologic laboratory and for both clinical and research utility. Originally, 1814 cellular images, divided into 4 diagnostic categories: superficial cell, intermediate cell, parabasal cell, dysplastic cell including low grade and high grade squamous intraepithelial lesion (LSIL and HSIL), from the liquid-based cervical smears with Papanicolaou stain, in 100,200 and 400 microscopic magnification were captured by digital camera. After review by pathologic experts, 503 images with peer agreement of cellular classification, were selected for further study. Automatically, self-designed computer program will analyze gray-level images by 13 morphometric parameters focusing on the objective measurement of nuclear size, nuclear shape, N/C ratio and chromatin pattern. Classification of signature patterns using a novel statistical support vector machine (SVM) based algorithm was created. Results show the selected morphometric parameters can be correctly differentiated the dysplastic cells from normal cells (statistic significance p<0.001). SVM classifier can achieve a high accuracy for cellular classification (accuracy up to 94.43%, sensitivity: 100% and specificity: 98.5%, no false negative cases). Conclusively, the proposed system can provide an inexpensive, feasible, accurate, quantitative tool for evaluating gynecologic specimens.
Index of Contents
中文摘要 I
Abstract III
誌謝 V
Index of Contents VII
List of Figures IX
List of Tables X
Chapter 1 Introduction 1
1.1 Background and Motivation 1
1.2 Objective 4
1.3 The Framework of Thesis 5
Chapter 2 Materials and Methods 6
2.1 Experimental Related Work 6
2.2 Acquisition of cell images 8
2.3 Cellular image filename processing 9
2.4 Image editing and processing 10
2.5 Objective micrometer image 12
2.6 PC-based cytopathologic image analysis (PCCIA) system 13
2.7 Support Vector Machine 13
Chapter 3 PC-based cytopathologic image analysis (PCCIA) system 17
3.1 Contour detection using circular seeded region growing method 18
3.2 Capture the texture with mask image 21
3.3 Measurement and analysis of cellular morphology and texture 22
3.3.1 The morphometric parameters 22
3.3.2 The texture parameters 24 Co-occurrence matrix 24 Tamura features 24
Chapter 4 Experiment and Discussion 27
4.1 Experiment sample 27
4.2 Evaluate the morphometric parameters 28
4.3 Classification of cell types using SVM 31
4.4 Results 35
Chapter 5 Conclusion and Future Work 40
5.1 Discussion and Conclusion 40
5.2 Future Work 43
Reference 43
Appendix 50
Arora B, Setia S, Rekhi B. 2006. Role of computerized morphometric analysis in diagnosis of effusion specimens. Diagn Cytopathol 34(10):670-5.
Bengtsson EW, Nordin B. 1993. Image analysis in cytology: DNA-histogramming versus cervical smear prescreening. Ann Biol Clin (Paris) 51(1):27-38.
Cancer Registry Annual Report, Republic of China. Executive, Yuan, ROC, Department of Health. "Cause of Mortality Report 2004," http://www.bhp.doh.gov.tw/BHP/index.jsp
Chang CC, Lin CJ, 2001. LIBSVM: a library for support vector machines. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm.
Chang CC, Lin CJ. 2001. Training nu-Support Vector Classifiers: Theory and Algorithms. Neural Computation 13(9): 2119-2147.
Cristianini N, Shawe-Taylor J 2000. An Introduction to Support Vector Machines and Other Kernel-Based Methods. Cambridge UK: Cambridge University Press.
Dawson AE. 2004. Can we change the way we screen?: the ThinPrep Imaging System. Cancer 102(6):340-4.
DeMay RM. 1997. Common problems in Papanicolaou smear interpretation. Arch Pathol Lab Med 121(3):229-38.
Doornewaard H, van der Schouw YT, van der Graaf Y, Bos AB, van den Tweel JG. 1999. Observer variation in cytologic grading for cervical dysplasia of Papanicolaou smears with the PAPNET testing system. Cancer 87(4):178-83.
Einstein AJ, Gil J, Wallenstein S, Bodian CA, Sanchez M, Burstein DE, Wu HS, Liu Z. 1997. Reproducibility and accuracy of interactive segmentation procedures for image analysis in cytology. J Microsc 188(Pt 2):136-48.
Grohs DH. 1998. Impact of automated technology on the cervical cytologic smear. A comparison of cost. Acta Cytol 42(1):165-70.
Harris MV, Cason Z, Benghuzzi H, Tucci M. 2000. Cytomorphological assessment of benign and malignant dense hyperchromatic groups in cervicovaginal smears. Biomed Sci Instrum 36:349-54.
Hartmann KE, Nanda K, Hall S, Myers E. 2001. Technologic advances for evaluation of cervical cytology: is newer better? Obstet Gynecol Surv 56(12):765-74.
Husain OA, Watts KC, Lorriman F, Butler B, Tucker J, Carothers A, Eason P, Farrow S, Rutovitz D, Stark M. 1993. Semi-automated cervical smear pre-screening systems: an evaluation of the Cytoscan-110. Anal Cell Pathol 5(1):49-68.
Idicula-Thomas S, Kulkarni AJ, Kulkarni BD, Jayaraman VK, Balaji PV. 2006. A support vector machine-based method for predicting the propensity of a protein to be soluble or to form inclusion body on overexpression in Escherichia coli. Bioinformatics 22(3):278-84.
Ikeguchi M, Sakatani T, Endo K, Makino M, Kaibara N. 1999. Computerized nuclear morphometry is a useful technique for evaluating the high metastatic potential of colorectal adenocarcinoma. Cancer 86(10):1944-51.
Kardos TF. 2004. The FocalPoint System: FocalPoint slide profiler and FocalPoint GS. Cancer 102(6):334-9.
Kasson PM, Huppa JB, Davis MM, Brunger AT. 2005. A hybrid machine-learning approach for segmentation of protein localization data. Bioinformatics 21(19):3778-86.
Kirillov V, Stebenyaeva E, Paplevka A, Demidchik E. 2006. A rapid method for diagnosing regional metastases of papillary thyroid cancer with morphometry. Microsc Res Tech 69(9):721-8.
Kirillov VA, Stebenyaeva EE, Paplevka AA, Demidchik EP. 2005. Quantitative changes in thyroid lymphoid cells as a marker of malignancy. Anal Quant Cytol Histol 27(2):101-10.
Kirillov VA, Yuschenko YP, Paplevka AA, Demidchik EP. 2001. Thyroid carcinoma diagnosis based on a set of karyometric parameters of follicular cells. Cancer 92(7):1818-27.
Ku NN. 1999. Automated Papanicolaou smear analysis as a screening tool for female lower genital tract malignancies. Curr Opin Obstet Gynecol 11(1):41-3
Linder J. 1997. Automation of the Papanicolaou smear: a technology assessment perspective. Arch Pathol Lab Med 121(3):282-6.
Lin HC, Chiu CY, Yang SN. 2003. Finding Textures by Textual Descriptions, Visual Examples, and Relevance Feedbacks. Pattern Recognition Letters 24: 2255-2267.
Mackin RW, Jr., Newton LM, Turner JN, Roysam B. 1998. Advances in high-speed, three-dimensional imaging and automated segmentation algorithms for thick and overlapped clusters in cytologic preparations. Application to cervical smears. Anal Quant Cytol Histol 20(2):105-21.
Mango LJ. 1996. Reducing false negatives in clinical practice: the role of neural network technology. Am J Obstet Gynecol 175(4 Pt 2):1114-9.
Mavroforakis ME, Georgiou HV, Dimitropoulos N, Cavouras D, Theodoridis S. 2006. Mammographic masses characterization based on localized texture and dataset fractal analysis using linear, neural and support vector machine classifiers. Artif Intell Med 37(2):145-62.
Murata S, Herman P, Iwashina M, Mochizuki K, Nakazawa T, Kondo T, Nakamura N, Lakowicz JR, Katoh R. 2005. Application of microscopic Forster resonance energy transfer to cytological diagnosis of the thyroid tumors. J Biomed Opt 10(3):034008.
Murata S, Mochizuki K, Nakazawa T, Kondo T, Nakamura N, Yamashita H, Urata Y, Ashihara T, Katoh R. 2002. Detection of underlying characteristics of nuclear chromatin patterns of thyroid tumor cells using texture and factor analyses. Cytometry 49(3):91-5.
Murata S, Mochizuki K, Nakazawa T, Kondo T, Nakamura N, Yamashita H, Urata Y, Ashihara T, Katoh R. 2003. Morphological abstraction of thyroid tumor cell nuclei using morphometry with factor analysis. Microsc Res Tech 61(5):457-62.
Nunobiki O, Sato M, Taniguchi E, Tang W, Nakamura M, Utsunomiya H, Nakamura Y, Mori I, Kakudo K. 2002. Color image analysis of cervical neoplasia using RGB computer color specification. Anal Quant Cytol Histol 24(5):289-94.
Naib ZM. 1996. Cytopathology. 4th ed. Bosten New York: Little, Brown and Company pp7-65.
Osterheld MC, Liette C, Anca M. 2005. Image cytometry: an aid for cytological diagnosis of pleural effusions. Diagn Cytopathol 32(3):173-6.
Pycock D, Taylor CJ. 1981. The magiscan image analyser as a diagnostic aid in cytology. Anal Quant Cytol 3(1):49-54.
Ramesh BV, Padaki VC, Hegde KS, Hazarika D, Verghese CA. 1992. An interactive image analysis system for quantitative cytology & to classify cervical cells. Indian J Med Res 96:338-43.
Stoler MH. 2000. Advances in cervical screening technology. Mod Pathol 13(3):275-84.
Tamura H, Mori S, Yamawaki T. 1978. Texture Features Corresponding to Visual Perception. IEEE Transactions on Systems, Man, and Cybernetics, Smc-8(6):460-473.
Taylor RN, Gagnon M, Lange J, Lee T, Draut R, Kujawski E. 1999. CytoView. A prototype computer image-based Papanicolaou smear proficiency test. Acta Cytol 43(6):1045-51.
Tsantis S, Cavouras D, Kalatzis I, Piliouras N, Dimitropoulos N, Nikiforidis G. 2005. Development of a support vector machine-based image analysis system for assessing the thyroid nodule malignancy risk on ultrasound. Ultrasound Med Biol 31(11):1451-9.
Tucker JH. 1976. CERVISCAN: an image analysis system for experiments in automatic cervical smear prescreening. Comput Biomed Res 9(2):93-107.
Theodoridis S, Koutroumbas K 2003. Pattern Recognition. 2nd ed., Academic Press, San Diego, USA.
Vapnik V. 1995. The nature of statistical learning theory. New York: Springer, 1st ed.
Wang SL, Wu MT, Yang SF, Chan HM, Chai CY. 2005. Computerized nuclear morphometry in thyroid follicular neoplasms. Pathol Int 55(11):703-6.
Yonekawa S, Sakai N, Kitani O. 1996. Identification of Idealized Leaf Types Using Simple Dimensionless Shape Factors by Image Analysis. Transactions of the ASAE, 39( 4): 1525-1533.
Zahniser DJ, Wong KL, Brenner JF, Ball HG, Garcia GL, Hutchinson ML. 1991. Contextual analysis and intermediate cell markers enhance high-resolution cell image analysis for automated cervical smear diagnosis. Cytometry 12(1):10-4.
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔