本篇論文提出一套應用色彩資訊於顯微影像，對腫瘤組織進行自動化分割的系統，簡稱為腫瘤細胞自動分割系統。此系統分析彩色影像並進行腫瘤組織分類，主要包含三個階段，第一階段是色彩標準化，可以降低同一個病人或是不同病人之間切片影像的差異程度，進而對同一期病患之組織切片影像，只取一組參數，就能適用於此期病患所有組織切片影像；第二階段是自動化特徵擷取，除了能有效節省人為選取樣本的人力及時間，更能減少人為手動圈選時所造成的誤差；第三階段是主成份分析，這部分是用來找出訓練影像中的色彩特徵，並計算影像中每個像素對各個特徵中心的馬式距離，最後將該像素分到馬式距離最小的那一類。之後，我們以半自動分割的結果作為ground truth，來衡量本篇所提出方法的正確性。實驗結果顯示兩者具有高度的一致性。因此，本文所提出的演算法可以有效分割口腔癌顯微影像，進而計算特徵參數作為診斷口腔癌之参考。另外，它亦能運用在所有以相同方法染色的切片顯微影像上。

This paper presents a new color-based approach that applies automatic segmentation to tumor tissue classification on microscopy images. The color-based image analysis for tissue classification consists of three stages: (1) the color normalization aimed at reducing the quality variation of tissue image within samples of each individual subject or across subjects; (2) the automatic sampling from tissue image to eliminate the manually done time consuming sampling work; and (3) principal component analysis (PCA) to characterize color features with a given set of training data. Then our system classifies every pixel to a cluster that has the minimal Mahalanobis distance between the cluster center and the corresponding pixel than all the other ones. We evaluate the algorithm by comparing the performance of the proposed method with the semi-automated one. Experimental studies show good consistency between the auto and semi-automatic methods. Therefore, the proposed algorithm provides an effective tool for the evaluation of oral cancer images, and it can also be applied to other microscopic images with the same type of tissue staining.

[1] ‘Control of Oral Cancer in Developing Countries. A W.H.O. meeting.’, Bulletin of the World Health Organization, vol. 62, no. 6, pp. 817-830, 1984.
[2] Chi-Shean Chan et al, Cancer registry annual report 1996, Taiwan, Republic of China, Department of Health, The Executive Yuan, 1996.
[3]http://ctpcweb1.cgmh.org.tw:8080/cancer/home.nsf/
5ffd30cee105d55648256b3b003c4249/f70b7af02562e
15748256cb40009ef88?OpenDocument
[4]http://prostatecancer.about.com/od/
newdiagnosishelp/a/cancerstaging.htm
[5] J. Folkman, ‘What is the evidence that tumors are angiogenesis dependent?’ Journal of the National Cancer Institute, vol. 82, no. 1, pp. 4-7, 1990.
[6] L.A. Liotta and J. Kleinerman et al, ‘Quantitative relationships of intravascular tumor-cells, tumor vessels, and pulmonary metastases following tumor implantation,’ Cancer Res, vol. 34, no. 5, pp. 997-1004, 1974.
[7] K. Axelsson and Britt-Marie E. Ljung et al, ‘Tumor angiogenesis as a prognostic assay for invasive ductal breast-carcinoma,’ Journal of the National Cancer Institute, vol. 87, no. 13, pp. 997-1008, 1995.
[8] A.W. Khan and A.P. Dhillon et al., ‘Prognostic significance of intratumoural microvessel density (IMD) in resected pancreatic and ampullary cancers to standard histopathological variables and survival,’ Eur J Surg Oncol, pp. 637-644, 2002.
[9] A. Gasinska and K. Urbanski et al., ‘Prognostic significance of intratumour microvessel density and haemoglobin level in carcinoma of the uterine cervix,’ Acta Oncol, vol. 41, no. 5, pp. 437-443, 2002.
[10] N. Ikeda and M. Adachi et al., ‘Prognostic significance of angiogenesis in human pancreatic cancer,’ British Journal of Cancer, vol. 79, no. 9, pp. 1553-1563, 1999.
[11] A. Rajan and J. Kusum et al, ‘Angiogenesis as an independent prognostic indicator in node-negative breast cancer,’ Anal Quant Cytol Histol, vol. 24, no. 4, pp. 228-233, 2002.
[12] W. Jung and J. Zhang, et al., ‘Advances in oral cancer detection using optical coherence tomography,’ IEEE Journal of Selected Topics in Quantum Electronics, vol. 11, no. 4, pp. 811-817, 2005.
[13] C. F. Jiang, C. Y. Wang, and C. P. Chiang, Oral cancer detection in fluorescent image by color image fusion. 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, vol. 2, pp. 1260-1262, 2004.
[14] R. Rodriguez, T. E. Alarcon and O. Pacheco, ‘A new strategy to obtain robust markers for blood vessels segmentation by using the watersheds method,’ Journal of Computers in Biology and Medicine, vol. 35, no.8, pp. 665-686, 2005.
[15] R. Rodriguez, ‘A strategy for blood vessels segmentation based on the threshold which combines statistical and scale space filter application to the study of angiogenesis,’ Computer Methods and Programs in Biomedicine, vol. 82, no. 1, pp. 1-9, 2006.
[16] C. G. Loukas and G. D. Wilson et al., ‘An image analysis-based approach for automated counting of cancer cell nuclei in tissue sections,’ Cytometry Part A, vol. 55, no. 1, pp. 30-42, 2003.
[17] C. G. Loukas and A. Linney, ‘On a relaxation-labelling algorithm for quantitative assessment of tumour vasculature in tissue section images,’ Computers in Biology and Medicine, vol. 35, no. 2, pp. 157-171, 2005.
[18] S. Cui and H. Hano, ‘Enhanced CD34 expression of sinusoid-like vascular endothelial cells in hepatocellular carcinoma,’ Pathology International, vol. 46, pp. 751-756, 1996.
[19] R.O. Duda, P.E. Hart and D.G. Stork, Pattern Classification Second Edition, WILEY INTERSCIENCE, 2001.
[20] E. Reinhard and M. Ashikhmin et al., ‘Color transfer between images,’ IEEE Computer Graphics and Applications, vol. 21, no. 5, pp. 34-41, 2001.
[21] Y.N. Sun and M.H. Horng, ‘Assessing liver tissue fibrosis with an automatic computer morphometry system,’ IEEE Engineering in Medicine and Biology, vol. 16, no. 3, pp. 66-73, 1997.
[22] E. Horowitz, S. Sahni and D. Mehta, FUNDAMENTALS OF DATA STRUCTURES IN C++, COMPUTER SCIENCE PRESS, 2001.
[23] R. A. Johnson, Probability and Statistics for Engineers 7th Edition, Pearson Prentice Hall, 2005.
[24] S. Sharma, APPLIED MULTIVARIATE TECHNIQUES, WILEY, pp. 375-378, 1996.
[25] L. L. He, Y. Z. Feng, Y. P. Gu, ‘Expression of Endostatin and MVD in Pancreatic Carcinomas’, Chinese Journal of Clinical and Experimental Pathology, vol. 21, no. 3, pp. 335-338, 2005.
[26] H. H. Su and S. T. Chu et al, ‘Spindle Cell Carcinoma of the Oral Cavity and Oropharynx: Factors Affecting Outcome’, J Chin Med Assoc, vol. 69, no. 10, pp.478-483, 2006.
[27] J. B. MacQueen, ‘Some Methods for classification and Analysis of Multivariate Observations’, Proceedings of 5-th Berkeley Symposium on Mathematical Statistics and Probability, vol. 1, pp.281-297, 1967.