跳到主要內容

臺灣博碩士論文加值系統

(3.235.56.11) 您好!臺灣時間:2021/07/29 11:40
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

: 
twitterline
研究生:陳俊宇
研究生(外文):CHEN, CHUN-YU
論文名稱:利用NullICA演算法於過度完備系統下之研究
論文名稱(外文):Overcomplete representation by NullICA algorithm
指導教授:汪群超
指導教授(外文):WANG, CHUN-CHAO
學位類別:碩士
校院名稱:國立臺北大學
系所名稱:統計學系
學門:數學及統計學門
學類:統計學類
論文種類:學術論文
論文出版年:2007
畢業學年度:95
語文別:中文
論文頁數:69
中文關鍵詞:獨立成份分析過度完備系統NullICA 演算法影像特徵擷取
外文關鍵詞:independent component analysisovercomplete representationNullICA algorithmimage feature extraction
相關次數:
  • 被引用被引用:0
  • 點閱點閱:115
  • 評分評分:
  • 下載下載:17
  • 收藏至我的研究室書目清單書目收藏:1
獨立成份分析(Independent Component Analysis, 簡稱ICA)是從1990開始發展的資訊處理技術,其主要目的是從觀察到的混合資料(Mixing data)中找出原始資料(Sources)。 ICA可依據混合矩陣A(Mxm)分成完備獨立成份分析(M=m)與過度完備獨立成份分析(M>m),而ICA實際應用大多是過度完備的情況,其混合矩陣為長方形矩陣,必須從低維度的混合資料找回原本高維度的原始資料,所以過度完備的ICA問題比完備的ICA還複雜許多。
本論文針對Chen(2003)所提出的NullICA方法,在原始資料為雙倍指數(Double Exponential, 簡稱DE)分配假設下,利用模擬資料及聲音資料於過度完備的ICA情況下分析NullICA的特性、效能與限制,最後成功將NullICA演算法應用在影像特徵擷取的影像處理上。
Independent component analysis (also called ICA) is an algorithm developed since 1990. The goal of ICA is to estimate the unknown mixing matrix A and to recover the original sources s given only the observations x in which x=As. According to the structure of the Mxm mixing matrix A, ICA can divided into complete ICA (M=m) and overcomplete ICA (M>m). In most ICA applications, the overcomplete situations are usually seen in which the mixing matrix A is a rectangular matrix with more columns than rows. It means that we must recover high dimensional sources from low dimensional observation data. In this way, the overcomplete ICA question is more complicated.
In this thesis, the NullICA algorithm (Chen 2003) was employed based on the assumption that the sources are distributed in a Double Exponential fashion. To analyze the characteristics, efficiency and the constraints of NullICA algorithm on the overcomplete ICA situation, both the simulated and real sound data are tested. Finally we successfully applied the NullICA algorithm on extracting basis features from image frames.
目錄

1:緒論 4
1.1研究背景介紹..........................................4
1.2文獻回顧..............................................9

2:NullICA演算法 12
2.1零核空間在過度完備線性系統下的表現形式....................12
2.2NullICA參數估計架構....................................13
2.3Langevin-EulerMoves演算法.............................15
2.4TheGivensSampler演算法................................16

3:NullICA演算法在原始資料為DE分配的情況下 19
3.1估計參數c.............................................19
3.2估計參數D.............................................20
3.3估計參數U與V...........................................23
3.4DE分配參數α的估計與分析.................................24

4:實驗結果 27
4.1模擬..................................................27
4.2盲蔽訊號源分離..........................................37
4.3NullICA演算法的特性、效能與限制..........................41
4.3.1觀察-假設混合矩陣A已知.................................41
4.3.2觀察-假設c固定已知....................................44
4.3.3觀察-假設參數D已知....................................49
4.4影像特徵擷取............................................50

5:研究成果及討論 57

附錄 58

參考文獻 68
[1] Bell, A. J. and Sejnowski, T. J. (1995). An information maximization approach to blind separation and blind deconvolution. Neural Computation, 7(6): 1129-1159.
[2] Bell, A. J. and Sejnowski, T. J. (1997). The ’Independent components’of natural scenes are edge filters. Vision Research, 37: 3327-3338.
[3] Casella George and Roger L. Berger.(2002). Statistical inference. Duxbury, U.S.A.
[4] Chen, R.-B. (2003). A Null-space Algorithm for Overcomplete Blind Source Separation. Ph.D. dissertation, Department of Statistics, University of California at Los Angeles.
[5] Chen, R.-B. and Chiang, S.-W. (2006). Overcomplete Blind Source Separation for Time-series Processes. Journal of the Chinese Statistical Association, 44: 342-363.
[6] Duane, S., Kennedy, A. D., Pendleton, B. J. and Roweth, D. (1987). Hybrid monte-carlo. Physics Letters B, 195(2): 216-222.
[7] Hastings, W. (1970). Monte carlo sampling methods using markov chains and their applications. Biometrika, 57: 97-109.
[8] Hyv¨arinen, A. and Oja. E. (1997). A fast fixed-point algorithm for independent component analysis. Neural Computation, 9(7): 1483-1492.
[9] Hyv¨arinen, A., Karhunen, J. and Oja. E. (2001). Independent component analysis. John Wiley, New York.
[10] Lee, T.-W., Lewicki, M. S., Girolami, M. and Sejnowski, T. J. (1999). Blind source separation of more sources then mixtures using overcomplete representation. IEEE Signal Processing Letters, 6(4): 87-90.
[11] Lewicki, M. S. and Olshausen, B. A. (1999). A probabilistic framework for the adaptation and comparison of image codes. J. Opt. Soc. Am. A: Optics, Image Science, and Vision, 16(7): 1587-1601.
[12] Lewicki, M. S. and Sejnowski, T. J. (2000). Learing overcomplete representations. Neural Computation, 12: 337-365.
[13] Linsker, R. (1992). Local synaptic learing rules suffice to maximize mutual information in a linear network. Neural Computation, 4: 691-702.
[14] Olshausen, B. A. (1996). Learning linear, sparse, factorial codes. A.I. Memo, 1580, Massachusetts Institute of Technology.
[15] Olshausen, B. A. and Field D. J. (1996). Emergence of simple-cell receptivefield properties by learning a sparse code for natural images. Nature, 381: 607-609.
[16] Olshausen, B. A. and Field D. J. (1997). Sparse coding with an overcomplete basis set: A strategy employed by V1? Vision Research, 11: 3311-3325.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top