臺灣博碩士論文加值系統

子宮頸癌是台灣女性癌症中排名前位的癌症，藉由子宮頸抹片檢查可以有效地做出預防治療。然而，人工檢查這些抹片影像是相當沒有效率的。電腦輔助的判斷系統除了減少人力消耗、增進效率，還能減少因疲勞造成的誤判問題。
本研究的目的為，將從子宮頸抹片影像中，精確切割出細胞質以及細胞核的區域。其中藉由細胞形狀及紋路等判定，將影像分類進行不同的切割流程，有效減少錯誤切割。
本篇論文使用金屬中心提供的細胞影像，最後利用五種不同的方法來評估本篇的切割效果。

Cervical cancer is the top cancer in Taiwan, and can be effectively prevented by cervical smear test. However, deal those smear images by human is inefficient. Computer-Assisted system can reducing human consumption, improve efficiency, but also reduce the fail decision caused by fatigue.
The aim of this study was to segment the cytoplasm and the nucleus from the cervical smear image. Which by the determination of cell shape, the image classification for different cutting process, effectively reducing the wrong cutting.
This paper uses the images provided by the metal center, and finally uses five different methods to evaluate the segmentation effect of this article.

摘要 i
Abstract ii
目錄 iii
圖目錄 vi
第一章 緒論 8
1.1 研究背景 8
1.2 研究動機與目的 9
1.3 論文架構 9
第二章 相關文獻探討 10
2.1 對比強化 10
2.1.1伽瑪校正 10
2.1.2 S曲線對比增強 11
2.2 切割方法 12
2.2.1主動輪廓模型(Active Contours Model(ACM)) 13
2.2.2水平集方法(Level set method)[11] 13
2.2.3 Chan-Vese方法(Chan-Vese method) 14
2.2.4 Distance Regularized Level Set Evolution (DRLSE) 15
第三章Hybrid-feature based cell segmentation (HFCS) method 17
3.1 整體流程 17
3.2.1 初步細胞切割 18
3.2.1.1取得灰階影像 19
3.2.1.2去除雜訊Tri-State Median filter 19
3.2.1.3去背 20
3.2.1.4斷開 21
3.2.1.5形狀判定 22
3.2.2 進階細胞切割 23
3.2.2.1梯度計算 24
3.2.2.2斷線連接 26
3.2.2.3小區合併 28
3.2.2.4細胞區域選定 29
3.2.2.5形狀判定 30
3.2.3 最終細胞切割 30
3.2.3.1去除暗區 32
3.2.3.2細胞切割 32
3.2.4 細胞核切割 33
3.2.4.1對比強化 34
3.2.4.2細胞核二值化 34
3.2.4.3去除雜點 35
3.2.4.4細胞核選定 36
第四章 實驗結果評估與比較 36
4.1 影像來源 36
4.2 實驗數據 38
4.3實驗評估與比較 40
第五章 結論與未來展望 45
參考文獻 46

[1]Wikipedia Foundation, Inc., Diabetes mellitus , Jul. 2017, https:// en.wikipedia.org/wiki/Cervical_cancer
[2]長庚紀念醫院曾志仁醫師，”子宮頸癌與子宮頸癌前期的防治” 中華民國衛生署子宮頸癌防治教材http://www.cgmh.org.tw/intr/intr5/c6700/Dr2/Work/Teaching/ Patient/子宮頸抹片衛教.htm.
[3]嚴孟祿、謝長堯，子宮頸癌在台灣，行政院衛生署80 年度委託研究計畫成果報告，民國九十年，台北。
[4]Tony F. Chan and Luminita A. Vese, “Active Contours Without Edges,” IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 10, NO. 2, FEBRUARY pp.266-277 2001
[5]Chunming Li, Chenyang Xu, Changfeng Gui, and Martin D. Fox, “Distance Regularized Level Set Evolution and ItsApplication to Image Segmentation,” IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 19, NO. 12, DECEMBER pp.3243-3254 2010
[6]NOBUYUKI OTSU , IEEE TRANSACTIONS ON SYSTREMS, MAN, AND CYBERNETICS, VOL. SMC-9, NO. 1 pp.62-66 1979.
[7]Tao Chen, Kai-Kuang Ma, and Li-Hui Chen, ” Tri-State Median Filter for Image Denoising,” IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 8, NO. 12, DECEMBER pp.1834-1838 1999
[8]Ce Li , Yang Yang ,n, Limei Xiao , Yongxin Li , Yannan Zhou , Jizhong Zhao, ”A novel image enhancement method using fuzzy Sure entropy,” Neurocomputing 215 pp.196–211 2016
[9]Wikipedia Foundation, Inc., Precision and recall , Nov. 2016, https://en.wikipedia.org/wiki/Precision_and_recall
[10]M. Kass, A. Witkin, and D. Terzopoulos, “Snakes: Active contour models,” Int. J. Comput. Vis., vol. 1, pp. 321–331, 1988.
[11]R. Malladi, J. A. Sethian, and B. C. Vemuri, “Shape modeling with front propagation: A level set approach,” IEEE Trans. Pattern Anal. Machine Intell., vol. 17, pp. 158–175, Feb. 1995.