臺灣博碩士論文加值系統

在人體中，淋巴結(Lymph Node, LN)是屬於內分泌系統中淋巴系統的一部分，通常也稱其為淋巴腺。身體中的每一個器官，幾乎都有淋巴結的存在，它的功能是幫助我們抵擋外來細菌及病毒的攻擊。正因如此，淋巴結的病變是很常見的，其中，淋巴轉移更是許多惡性腫瘤用來分級的依據。超音波影像是常見用來診斷淋巴結是否病變的工具之一。臨床醫學上，淋巴病變通常結合專業醫師的經驗和臨床組織中的病理切片來診斷，但缺點是檢測非常耗時。因此，本論文將提出一個淋巴結超音波影像的疾病分類系統，藉由特徵擷取方法，在影像中取得其紋理特徵，並且結合粒子群體最佳化神經網路(Particle Swarm Optimization neural network, PSONN)與波茲曼機率函數做為最佳化特徵選取，最後再利用支援向量機(Support Vector Machine, SVM)的分類技術，進而有效地達到特徵維度的降低和正確分類兩個目的。實驗結果顯示，經由篩選出來的特徵確實能達到維度降低和正確分類的效果，同時也縮短了計算的時間。

A lymph node (LN) is a part of the lymphatic system that exists in human body and every apparatus. LN can resist virus and germs. There are many kinds of pathological change in LN. Metastatic is one of the important indexes in staging malignant tumors. One convenient tool to observe LN is the use of an ultrasonic image. Clinical physicians judge a nosology by pathological section and experience of the professionals. Shortcoming of this method is that it requires lots of precious time of clinical physicians. In engineer’s view, we can help with some technology to classify images took with ultrasound. In this paper, we propose a system that classifies Lymph Node with different pathological change in ultrasonic images. Features are selected as well as extracted from the ultrasonic images. Furthermore, a feature-selecting method that integrates the particle swarm optimization neural network (PSONN) with Boltzmann probabilistic and the support vector machine (SVM) neural network is adopted to classify these images. The experimental results show that the proposed approach decreases the number of the selected features and achieves a high accuracy in classification.

中文摘要 ---------------------------------------------------------------------- i
Abstract ---------------------------------------------------------------------- ii
目錄 ---------------------------------------------------------------------- iii
圖片索引 ---------------------------------------------------------------------- v
表格索引 ---------------------------------------------------------------------- vii
第一章 緒論 -------------------------------------------------------------- 1
1.1 研究動機 -------------------------------------------------------- 1
1.2 超音波淋巴結影像簡介 -------------------------------------- 1
1.3 相關研究 -------------------------------------------------------- 3
1.4 研究方法 -------------------------------------------------------- 4
1.5 章節大綱 -------------------------------------------------------- 6
第二章 相關理論 -------------------------------------------------------- 7
2.1 特徵擷取 -------------------------------------------------------- 7
2.1.1 空間性灰階共生矩陣 ----------------------------------------- 8
2.1.2 統計特徵矩陣 -------------------------------------------------- 12
2.1.3 灰階長度紋理矩陣 -------------------------------------------- 12
2.1.4 LAW紋理能量矩陣 ------------------------------------------ 14
2.1.5 灰階相依紋理矩陣 -------------------------------------------- 16
2.1.6 小波特徵 -------------------------------------------------------- 17
2.1.7 區域傅立葉係數特徵 ----------------------------------------- 18
2.2 特徵選取 -------------------------------------------------------- 18
2.2.1 基因演算法 ----------------------------------------------------- 19
2.2.2 循序前進浮動搜尋演算法 ----------------------------------- 21
2.2.3 粒子群體最佳化演算法 -------------------------------------- 23
2.3 波茲曼機 -------------------------------------------------------- 25
2.4 支援向量機 ----------------------------------------------------- 26
第三章 研究方法 -------------------------------------------------------- 33
3.1 ROI區域影像 -------------------------------------------------- 34
3.2 結合粒子群體最佳化神經網路(PSONN)與波茲曼函數之特徵選取 -----------------------------------------------------
35
3.2.1 基因表示法 ----------------------------------------------------- 36
3.2.2 粒子群體最佳化神經網路(PSONN) --------------------- 37
3.3 網路分類 -------------------------------------------------------- 41
3.3.1 一對一分類模式 ----------------------------------------------- 42
3.3.2 一對多分類模式 ----------------------------------------------- 43
第四章 實驗結果與討論 ----------------------------------------------- 45
4.1 資料來源 -------------------------------------------------------- 45
4.2 特徵選取結果 -------------------------------------------------- 48
4.2.1 探討不同參數的組合對整體效能的影響 ----------------- 50
4.2.2 傳統PSO演算法和PSONN之比較 ---------------------- 55
4.2.3 不同特徵選取法之比較 -------------------------------------- 56
第五章 結論 -------------------------------------------------------------- 59
參考文獻 --------------------------------------------------------------------- 60

[1]Il-Seok Oh, Jin-Seon Lee, Byung-Ro Moon, “Hybrid genetic algorithms for feature selection”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 26, pp. 1424-1437, November 2004.
[2]Stavroula G. Mougiakakou, Ioannis K. Valavanis, Alexandra Nikita, Konstantina S. Nikita, “Differential diagnosis of CT focal liver lesions using texture features, feature selection and ensemble driven classifiers”, Artificial Intelligence in Medicine, vol. 41, pp. 25-37, September 2007.
[3]Cheng-Lung Huang, Chieh-Jen Wang, “A GA-based feature selection and parameters optimization for support vector machines”, Expert Systems with Applications, vol. 31, pp. 231-240, 2006.
[4]J. Jiang, B. Yao, A.M Wason, ”A genetic algorithm design for microcalcification detection and classification in digital mammograms”, Computerized Medical Imaging and Graphics, vol. 31, pp. 49-61, 2007.
[5]Chin-Chin Lai, Chuan-Yu Chang, ”A hierarchical evolutionary algorithm for automatic medical image segmentation”, Expert Systems with Applications, vol. 38, no. 3, 2007.
[6]Hsuan-Ming Feng, “Self-generation RBFNs using evolutional PSO learning”, Neurocomputing, vol. 70, pp. 241-251, 2006.
[7]Li-Yeh Chuang, Hsueh-Wei Chang, Chung-Jui Tu, Cheng-Hong Yang, “Improved binary PSO for feature selection using gene expression data”, Computational Biology and Chemistry, vol. 32, pp. 29-38, 2008.
[8]James Kennddy, Russell C. Eberhart, “A discrete binary version of the particle swarm algorithm”, IEEE Conference on Neural Networks, pp. 1942-1948, 1995.
[9]Haralick. RM, Shanugam K, and Dinstein I, “Textural features for image classification”, IEEE Transactions on Systems, Man and Cybernetics, vol. 3, No, pp. 610-621, 1973.
[10]Chung-Ming Wu, Yung-Chang Chen, “Statistical feature matrix for texture analysis”, Proceeding of CVGIP Conference on Graphical Models and Image Processing, vol. 54, pp. 407-419, 1992.
[11]Galloway, M.M., “Texture Analysis Using Gray Run Lengths”, Computer Graphics and Image Processing, Vol. 4, pp. 172-179, 1975.
[12]Chengjun Sun and William G. Wee, “Neighboring gray level dependence matrix for texture classification”, Computer Vision, Graphics, and Image Processing, Vol. 23, No. 3, pp. 341-352, 1983.
[13]K. Laws, “Textures image segmentation”, Ph.D. Dissertation, University of Southern California, 1980.
[14]Antonini, M. Barlaud, M. Mathieu, P. and Daubechies, I, “Image coding using wavelet transform”, IEEE Transactions on Image Processing, Vol. 1, No. 2, pp. 205-220, 1992.
[15]Rafael C. Gonzalez, Richard E. Woods, “Digital Image Processing”, Prentice- Hall Internation, Inc, 1998.
[16]Feng Zhou, Ju Fu Feng, and Qing Yun Shi, “Texture feature based on local Fourier transform”, Proceeding of International Conference on Image Processing, Vol. 2, pp. 610-613, 2001.
[17]Feng Zhou, Ju-Fu Feng, Qing Yun Shi, “Texture feature based on local fourier transform”, Proceeding of International conference on Image Processing, vol.2, pp.610-613, 2001.
[18]Simon Haykin, “Neural Networks”, Prentice-Hall Internation, 1999.
[19]Bernhard E. Boser, Isabelle M. Guyon, Vladimir N. Vapnik, “A training algorithm for optimal margin classifiers”, In Proceedings of the Fifth Annual Workshop on Computational Learning Theory, pages 144-152. ACM Press, 1992.
[20]Chih-Chung Chang and Chih-Jen Lin, LIBSVM: a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm.
[21]J. H. Holland, “Adaptation in natural and artificial systems”, The MIT Press, 1992.
[22]P. Pudil, F. J. Ferri, J. Novovicova, J. Kittler, “Floating Search Methods for Feature Selection with Nonmonotonic Criterion Functions”, Proceedings of the IAPR International Conference on Pattern Recognition, Vol. 2, pp. 279-283, 1994.
[23]Isabelle Guyon, Steve Gunn, Masoud Nikravesh, and Lotfi A. Zadeh, Feature Extraction: Foundations and Applications, Spriner, 2006.
[24]Chin-Wei Hsu, Chin. Jen. Lin, “A Comparison of Methods for Multiclass Support Vector Machines”, IEEE Transaction on Neural Networks, vol. 13, pp. 1026–1027, 2002.
[25]Ryan. Rifkin and Aldebaro. Klautau, “In Defense of One-vs-All Classification”, Journal of Machine Learning Research, Vol. 12, pp 101-141, 2004.
[26]LOGIQ 700 Expert Series: Conformance Statement for DICOM V3.0. Direction 2243479-100, Revision 1. Technical Publications. GE Healthcare, Chalfant St. Giles, UK. Available at: http://www.gehealthcare.com/usen/interoperability/ dicom/docs/logiq700_rev2.pdf
[27]葉怡成,”類神經網路模式應用與實作”,儒林圖書公司,2004.
[28]張斐章,張麗秋,”類神經網路”,東華書局,民94年.
[29]莊麗月, 楊正宏, 杜崇睿, 張學偉, “GA-PSO於基因表現資料的特徵選取”, 數位科技與創新管理國際研討會, 2006.
[30]張鴻豪, 使用支援向量委員會機器於適應性色彩空間切換之臉部追蹤, 國立雲林科技大學電子工程研究所碩士論文, 民95年.