Please use this identifier to cite or link to this item:
標題: 有效運用ICA與分類器於腦部MR影像顱內組織量化測量之探討
Exploring the Effectiveness of ICA with Classifier in Quantitative Volume Measurement of Brain MRI
作者: 王士偉
Wang, Shih-Wei
關鍵字: 獨立成分分析;Independent component analysis (ICA);盲信號分離;完備性獨立成分分析;支援向量機;擷取顱內組織;Blind source separation (BBS);Over-complete ICA (OC-ICA);Support vector machine (SVM);Skull-stripping.
出版社: 電機工程學系所
引用: [1] A. Hyvärinen, J. Karhunen and E. Oja, Independent Component Analysis, John Wiley & Sons, 2001. [2] Nakai, T., Muraki, S., Bagarinao, E., Miki Y., Takehara, Y., Matsuo, K., Kato, C., Sakahara, H., Isoda, Application of independent component analysis to magnetic resonance imaging for enhancing the contrast of gray and white matter, NeuroImage, vol. 21, pp. 251-260, 2004. [3] A.J. Bell and T.J. Sejnowski, An information-maximization approach to blind separation and blind deconvolution. Neural Computation, vol. 7, pp. 1129-1159, 1995. [4] A. Hyvärinen and E. Oja, A fast fixed-point for independent component analysis, Neural Computation, vol. 9, no. 7, pp. 1483-1492, 1997. [5] Athanasios Papoulis, Probability, Random Variables, and Stochastic Processes, McGraw-Hill, 3rd edition, 1991. [6]. A. Hyvärinen, J. Karhunen and E, Independent Component Analysis, Simon Haykin, Series Editor. [7] A. Hyvärinen, The fixed-point algorithm and maximum likelihood estimation for independent component analysis, Neural Processing Letters, 1999. To appear. [8] A. Hyvärinen, Survey on Independent Component Analysis, Neural Computing Surveys 2, 94-128, 1999. [9] S. Haykin, Neural Networks: A Comprehensive Foundation, 2nd ed., Prentice-Hall, 1999, Chapter 6. [10] V.N. Vapnik, Statistical Learning Theory, John Wiley & Sons, 1998. [11] Digabel, H., and Lantuejoul, C. Iterative algorithms, Actes du Second Symposium Europeen d''Analyse Quantitative des Microstructures en Sciences des Materiaux, Biologie et Medecine, Caen, 4-7 October 1977, J.-L. Chermant, Ed., Riederer Verlag, Stuttgart, pp. 85-99, 1978. [12] Lantuejoul, C. La squelettisation et son application aux mesures topologiques des mosaiques polycristallines. PhD thesis, Ecole des Mines, Paris, 1978. [13] 徐賢鈞,自適影像分割技術及三維重建-以腦部磁振造影解剖影像之大腦組織結構分割為案例,大葉大學工業工程學系碩士論文,民國九十三年六月。 [14] Hahn H.K. and Peitgen H.O., The skull striping problem in MRI solved by a single 3D watershed transform, Proc. MICCAL, LNCS 1935, pp.134-143, 2000. [15] Paul Tofts. Quantitative MRI of the Brain: Measuring Changes Caused by Disease, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England. [16] Sergios Theodoridis and Konstantinos Koutroumbas. Pattern Recognition, 2nd ed, Elsevier Science. [17]
由於獨立成分分析(ICA)在盲信號分離能力,它在核磁共振影像(MRI)分析中也具有相當好的表現。一般而言,核磁共振影像是使用完備性獨立成分分析(over-complete -ICA)來處理,也就是影像張數少於訊號來源數的盲點分離分析。除了利用完備性獨立成分分析結合支援向量機(SVM)來解決超過一個成分以上的腦部組織可能會被分離並被迫變成一個單一獨立成分(IC)的問題,本論文在上述方法中更進一步提出擷取顱內組織的前處理,使分類的效果更為顯著。為了證明我們所提出的方法,將在量化測量方面進行效能分析和評估。實驗結果顯示,在腦部磁振影像中擷取顱內組織的步驟在分析上的確是一個相當重要的前處理。

Independent component analysis (ICA) has shown promise in magnetic resonance (MR) image analysis due to its ability in blind source separation (BBS). The ICA used for MR image is generally over-complete in the sense that the number of images is usually less than the number of signal sources to be blindly separated. In addition to implement the over-complete ICA in conjunction with classification, namely support vector machine (SVM), to resolve that more than one brain tissue clusters may be separated and forced into a single independent component (IC), this thesis further presents a pre-processing stage, called skull-stripping, to remove non-brain tissue in an attempt to achieve significantly better classification in the ICA in conjunction with SVM. In order to demonstrate the proposed approach, we carry out quantitative measurement for performance analysis and evaluation. Experimental results show that skull-stripping in the brain MR image is one of the most important pre-processing stages for the analysis of spatial distribution.
其他識別: U0005-2507200713120800
Appears in Collections:電機工程學系所

Show full item record

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.