Skip navigation

DSpace

機構典藏 DSpace 系統致力於保存各式數位資料(如:文字、圖片、PDF)並使其易於取用。

點此認識 DSpace
DSpace logo
English
中文
  • 瀏覽論文
    • 校院系所
    • 出版年
    • 作者
    • 標題
    • 關鍵字
    • 指導教授
  • 搜尋 TDR
  • 授權 Q&A
    • 我的頁面
    • 接受 E-mail 通知
    • 編輯個人資料
  1. NTU Theses and Dissertations Repository
  2. 理學院
  3. 數學系
請用此 Handle URI 來引用此文件: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/23740
完整後設資料紀錄
DC 欄位值語言
dc.contributor.advisor陳素雲(Su-Yun Huang)
dc.contributor.authorPei-Shien Wuen
dc.contributor.author吳佩勳zh_TW
dc.date.accessioned2021-06-08T05:09:33Z-
dc.date.copyright2011-08-03
dc.date.issued2011
dc.date.submitted2011-07-23
dc.identifier.citationAnderson, T. W. (1963). Asymptotic theory for principal component analysis. Annals of Mathematical Statistics, 34, 122-148.
Carroll, J.D. and Chang, J. (1970). Analysis of individual di fferences in multidimensional scaling via an N-way generalization of 'Eckart-Young' decomposition. Psychometrika, 35, 283-319.
Chen, T.L. and Shiu, S.Y. (2007). A new clustering algorithm based on self-updating process. 2007 Proceedings of the American Statistical Association, Statistical Computing Section [CD-ROM], Salt Lake City, Utah.
DeRosier, D. and Klug, A. (1968). A reconstruction of 3-dimensional structure from electron micrographs. Nature, 217:130-134.
De Lathauwer, L., De Moor, B. and Vandewalle, J. (2000a). A multilinear singular value decomposition. SIAM J. Matrix Anal. Appl., 21, 1253-1278.
De Lathauwer, L., De Moor, B. and Vandewalle, J. (2000b). On the best rank-1 and rank-(R1,R2,...RN)approximation of higher-order tensors. SIAM J. Matrix Anal. Appl., 21, 1324-1342.
Fine, J. (1987). On the validity of the perturbation method in asymptotic theory. Statistics, 18, 401-414.
Harshman, R.A. (1970). Foundations of the PARAFAC procedure: Model and conditions for an 'explanatory' multi-mode factor analysis. UCLA Working Papers in Phonetics, 16, 1-84.
Henderson, H. V. and Searle, S. R. (1979). Vec and vech operators for matrices, with some uses in Jacobians and multivariate statistics.
Canadian J. Statistics, 7, 65-81. Jolli e, I.T. (2002). Principal Component Analysis. Springer, New York.
Kolda, T.G. and Bader, B.W. (2009). Tensor decompositions and applications. SIAM Review, 51(3), 455-500.
Li, B., Kim, M.K. and Altman, N. (2010). On dimension folding of matrix or array-valued statistical objects. Annals of Statistics, 38, 1094-1121.
Lu, H., Plataniotis, K. N. and Venetsanopoulos, A. N. (2008). MPCA: Multilinear principal component analysis of tensor objects. IEEE Transactions on Neural Networks, 19, 18-39.
Magnus, J. R. and Neudecker, H. (1979). The commutation matrix: some properties and applications. Annals of Statistics, 7, 381-394.
Sibson, R. (1979). Studies in the robustness of multidimensional scaling: perturbational analysis of classical scaling. J. Roy. Statist. Soc., 41,
217-229.
Tucker , L.R. (1966). Some mathematical notes of three-mode factor analysis.Psychometrika, 31, 279-311.
Tu, I. P., Chen H. and Chen, X. (2009). An eigenvector variability plot. Statistica Sinica, 19, 1741-1754.
Tyler, D. E. (1981). Asymptotic inference for eigenvectors. Annals of Statistics, 9, 725-736.
Yang, J., Zhang, D., Frangi, A.F. and Yang, J.Y. (2004). Two-dimensional PCA: a new approach to appearance-based face representation and recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 26, 131-137.
Ye, J. (2005). Generalized low rank approximations of matrices. Machine Learning, 61, 167-191.
Zhang, D. and Zhou, Z. H. (2005). (2D)2PCA: Two-directional two-dimensional PCA for e cient face representation and recognition. Neurocomputing, 69, 224-231. 44
dc.identifier.urihttp://tdr.lib.ntu.edu.tw/jspui/handle/123456789/23740-
dc.description.abstract在本論文中,我們將介绍多重線性的主要成分分析(MPCA),並推導出它在統計上的一些性質,之後將它運用在低溫電子顯微鏡的影像分析上。 由於此影像具有較低的雜訊比(SNR),因此將相類似的圖像做平均是必要的。 K-means是將樣本點做分類最常用的演算法。 然而,我們發現此方法對具有低雜訊比特性的低溫電子顯微鏡影像的分析是不理想的,因此這裡我們將採用self-updating process (SUP) 演算法 (Chen and Shiu, 2007) 運用此演算配合多重線性的主要成分分析來重建物體的結構。zh_TW
dc.description.abstractIn statistics, dimension reduction is a process of reducing the number of random variables under consideration, and can be divided into feature selection and feature extraction. Principal component analysis (PCA) belongs to the latter
category. Traditional linear techniques for dimensionality reduction like PCA reshapes image matrices into vectors. It leads to vectors in a very high-dimensional space and thus easily suff ers from the curse of dimensionality. Multilinear principal component analysis (MPCA) has the potential to serve the similar purpose for analyzing tensor structure data. MPCA aims to preserve the natural data structure, based on 2D matrices rather than 1D vectors, and searches for low-dimensional multilinear projections. It can decrease the dimensionality in a more stable and efficient way than traditional PCA. MPCA and other tensor decomposition methods have been shown to have good performance in both real data analysis and simulations (Ye, 2005; Lu, Plataniotis and Venetsanopoulos, 2008; Kolda and Bader, 2009; Li, Kim and Altman, 2010). However, there is not much statistical theoretic study of it. In this thesis, we place the MPCA in a statistical framework and investigate its statistical properties, including asymptotic distributions for principal components, associated projections and explained variances. We also apply it to electron microscopy images analysis. Due to the nature of low signal to noise ratio (SNR) of electron microscopy images, an averaging process for similar images is needed for denoising. The k-means algorithm is probably the most commonly used algorithm for clustering. However, we find it not ideal for low SNR electron microscopy images. The k-means algorithm needs quite some manual tuning and care in order to get reasonable clustering results. Here we adopt a self-updating process (SUP) clustering algorithm (Chen and Shiu, 2007) on the MPCA-extracted core tensors to recover the hidden cluster structure.
en
dc.description.provenanceMade available in DSpace on 2021-06-08T05:09:33Z (GMT). No. of bitstreams: 1
ntu-100-R98221038-1.pdf: 1242354 bytes, checksum: 8f16338659ce0f44f85664434c767ec4 (MD5)
Previous issue date: 2011
en
dc.description.tableofcontentsContents
致謝 i
中文摘要 ii
Abstract iii
Contents iv
List of Figures v
List of Tables v
1 Introduction and a motivating example 1
2 Literature review 6
2.1 Review of multilinear algebra . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 High order SVD (HOSVD) . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Generalized low rank approximations of matrices . . . . . . . . . . . . . . . 13
2.4 Two-dimensional PCA and (2D)2 PCA . . . . . . . . . . . . . . . . . . . . . 14
3 Multilinear PCA in statistical framework 17
3.1 Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Selection of dimensionality . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4 Asymptotic properties for MPCA 26
4.1 Asymptotic distributions for principal components, projections, cumulative
variance and explained variance . . . . . . . . . . . . . . . . . . . . . . . . . 27
5 Application to cryo-EM Images 31
5.1 Simulation data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
6 Conclusion 41
7 References 42
8 Appendix 45
List of Figures
1 Image information . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3 Slices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4 Mode-n fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5 Unfolded matrix example . . . . . . . . . . . . . . . . . . . . . 9
6 SVD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
7 HOSVD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
8 Flowchart of the cryo-EM image analysis process . . . . . . . 31
9 Projection-slice theorem . . . . . . . . . . . . . . . . . . . . . 32
10 Simulation data tensor representation . . . . . . . . . . . . . . 34
11 Conventional PCA . . . . . . . . . . . . . . . . . . . . . . . . 37
12 Multilinear PCA by GLRAM . . . . . . . . . . . . . . . . . . 38
13 The number of class variates by choosing different parameters 40
List of Tables
1 GLRAM algorithm (Ye, 2005) . . . . . . . . . . . . . . . . . . 22
2 SUP clustering algorithm (Chen and Shiu, 2007) . . . . . . . . 36
3 Methods comparison . . . . . . . . . . . . . . . . . . . . . . . 38
4 Correctness distributions for different methods . . . . . . . . . 39
dc.language.isoen
dc.title二階張量多維線性主成分分析在低溫電子顯微鏡影像的應用zh_TW
dc.titleOn Multilinear Principal Component Analysis of Order-Two Tensors With Application to Electron Microscopy Imagesen
dc.typeThesis
dc.date.schoolyear99-2
dc.description.degree碩士
dc.contributor.coadvisor杜憶萍(I-Ping Tu)
dc.contributor.oralexamcommittee陳宏(Hung Chen),王偉仲(Wei-Chung Wang),章為皓(Wei-Hau Chang)
dc.subject.keyword多維線性主成分分析,低溫電子顯微鏡影像,zh_TW
dc.subject.keywordMultilinear Principal Component Analysis,Electron Microscopy Images,en
dc.relation.page50
dc.rights.note未授權
dc.date.accepted2011-07-25
dc.contributor.author-college理學院zh_TW
dc.contributor.author-dept數學研究所zh_TW
顯示於系所單位:數學系

文件中的檔案:
檔案 大小格式 
ntu-100-1.pdf
  未授權公開取用
1.21 MBAdobe PDF
顯示文件簡單紀錄


系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。

社群連結
聯絡資訊
10617臺北市大安區羅斯福路四段1號
No.1 Sec.4, Roosevelt Rd., Taipei, Taiwan, R.O.C. 106
Tel: (02)33662353
Email: ntuetds@ntu.edu.tw
意見箱
相關連結
館藏目錄
國內圖書館整合查詢 MetaCat
臺大學術典藏 NTU Scholars
臺大圖書館數位典藏館
本站聲明
© NTU Library All Rights Reserved