本研究提出以相對次數(Relative frequency)分配為基之獨立成份分析(Independent component analysis, ICA) 模式，利用相對次數分配估計出聯合機率密度函數及邊際機率密度函數做為量化獨立性的量測準則，並結合粒子群最佳化(Particle Swarm Optimization, PSO)演算法的方法來搜尋最佳解所建立的獨立成分分析模式，可以還原具相關性的原始訊號。傳統上獨立成份分析法是以訊號資料的非高斯特性(Non-gauissianity) 來量測其獨立性，利用非高斯特性建立的獨立成分分析模式來還原混合訊號，原始訊號間的相關性也會被去除，因此使用傳統的獨立成份分析法無法有效還原具高相關性之訊號。由於本研究方法之聯合機率密度函數在訊號個數大於或等於3時的估算不穩定，目前僅適合訊號個數2筆之應用。

本研究所發展之獨立成份分析模式，主要應用於瑕疵檢測 (Defect detection)與動態影像偵測 (Motion detection)。在瑕疵檢測方面乃以TFT-LCD液晶面板為討論對象，利用LCD面板區域元件皆呈現週期排列之特徵，與本研究之獨立成份分析模式可還原具有高相關性之原始訊號的特性，發展出一個不需對檢測影像做定位之自我影像(Self-reference)檢測的機器視覺演算法。而本研究於動態影像偵測的應用上，主要是針對室內固定位置之影像，利用動態物體與靜態背景互相獨立的特徵，經由本研究之獨立成份分析模式，發展出一個對於光源變化有較高的抵抗性，且可大量減少運算時間而達到及時偵測出前景目標物的演算法。

In this study, an independent component analysis (ICA) model that directly measures the difference of the joint probability density function (p.d.f.) and the product of marginal p.d.fs is proposed. The p.d.fs are estimated from the relative frequency distributions, and the particle Swarm Optimization (PSO) algorithm is used to search for the best solution in the ICA model. The proposed ICA model can separate highly correlated data, which is not achievable by the well-known FastICA algorithm that uses non-gaussianity as the independency measure. Since the p.d.f. estimated in the ICA model is simply based on the count of relative frequency, it can only well separate mixture of two source signals. When it comes to more than two signals, the p.d.f. estimation performs unreliably. Therefore, the applications of the proposed ICA model in this research are restricted to the separation of two sources.

The proposed ICA model is applied to defect detection and motion detection, where the underlying signals show high correlation. For defect detection, panel surfaces of TFT-LCD and color filter are the main targets of study. In a panel image, each scan line shows a periodical pattern. By dividing a scan line into two segments of equal length, the two segments are only different by their translations. The proposed ICA model is applied to filter translation changes. A cross correlation-based similarity measure can then be used to identify anomalies in the inspection surface. For motion detection, the proposed ICA model is applied to separate foreground objects from the stationary background. The implementation of the proposed method is computationally fast, and is insensitive to illumination changes. Experimental results have shown that the proposed methods are very efficient and effective for defect detection and motion detection applications.

目錄
中文摘要 I
英文摘要 II
誌謝 III
目錄 IV
圖目錄 VI
表目錄 VIII
第一章 緒論 1
1.1研究背景與動機 1
1.2研究範疇及目的 2
1.3研究方法簡介 4
1.4章節架構說明 5
第二章 文獻回顧 6
2.1獨立成份分析法之獨立性量測 6
2.1.1峰態 ( KURTOSIS ) 7
2.1.2負熵 ( Negentropy ) 8
2.1.3共同資訊( Mutual information ) 9
2.1.4 核心密度函數估測 (Kernel density estimation) 11
2.2 快速獨立成份分析法( FASTICA ) 12
2.2.1 FastICA之簡介 12
2.2.2 Fast ICA演算流程 15
2.3獨立成份分析法之應用 16
2.3.1醫學訊號之解讀 16
2.3.2紋路分析 17
2.3.3影像復原 18
2.3.4人臉辨識 18
2.3.5瑕疵檢測 19
第三章 研究方法 20
3.1研究方法流程概述 20
3.2獨立成份分析(ICA) 22
3.2.1 ICA模型 24
3.2.2獨立性量測與機率密度函數之估算 25
3.3粒子群最佳化(PSO) 32
3.3.1粒子群最佳化之發展背景 32
3.3.2粒子群最佳化之概念說明 33
3.4 PSO-ICA相對於目前常用獨立成份分析法之特性 39
3.4.1 PSO-ICA演算法 39
3.4.2 PSO-ICA相對於快速獨立成份分析法(FastICA)之獨特性 43
第四章 獨立成份分析法於TFT-LCD表面瑕疵檢測之應用 55
4.1 現有之LCD瑕疵檢測技術 55
4.2瑕疵檢測之應用的研究目的與流程 58
4.2.1研究目的與優點 58
4.2.2研究流程步驟 60
4.3 初步實驗結果 71
4.3.1 系統架構與實驗環境 71
4.3.2 影響MVA-LCD瑕疵檢測結果之參數討論 72
4.3.3 本研究方法與FastICA於TFT-LCD檢測結果之比較 81
4.4 彩色濾光片實驗 84
第五章 獨立成份分析法於動態影像偵測之應用 111
5.1 動態影像偵測之相關文獻 111
5.2動態影像偵測的研究目的與研究流程 114
5.2.1研究目的與優點 114
5.2.2 動態影像偵測之流程步驟 114
5.3 初步實驗結果 126
5.3.1 系統架構與實驗環境 126
5.3.2 影響動態影像偵測結果之參數討論 127
5.3.3 本研究方法與FastICA、直接相減法結果之比較 135
5.4 背景變動的動態影像偵測 153
5.4.1 背景中窗簾的變動 153
5.4.2 背景中門開關的變動 165
第六章 結論及未來發展方向 178
參考文獻 180
附錄A 原始訊號間有相關性時的各個數據圖 184
附錄B 本研究方法與FASTICA於MVA-LCD測試其他結構一維影像結果 194
附錄C 不變動反混合矩陣而變動背景影像之結果 198
附錄D PSO搜尋次數與收斂之關係 201
附錄E程式說明 203

白清錫，1986年，「品質管制之統計方法」，第四版，中華民國品質管制學會發行，pp.18~20、pp.29~27。

錢志豪，2002年，「建構液晶顯示器(LCD)色彩偏差瑕疵之自動化視覺檢測系統之探討」，碩士論文，私立朝陽科技大學工業工與管理程研究所。

劉國慶，2003年，「利用磁振造影的血流灌注影像分割技術來評估腦部疾病」，碩士論文，國立陽明大學放射醫學科學研究所。

洪至懿，2003年，「特徵擷取與分類應用與想像左右手手指運動之腦波辨識」，碩士論文，國立陽明大學放射醫學科學研究所。

曾彥馨，2003年，「應用機器視覺於TFT面板之表面瑕疵檢測與分類」，碩士論文，私立元智大學工業與管理工程研究所。

林品杰，2005年，「應用獨立成份分析(ICA)濾波器於背光板與TFT-LCD面板之瑕疵檢測」，碩士論文，私立元智大學工業與管理工程研究所。

Bartlett, M. S., J. R. Movellan and T. J. Sejnowski, 2002, “Face recognition by independent component analysis,’’ IEEE Transactions on Neural Networks, Vol. 13, pp. 1450-1464.

Boscolo R., H. Pan, P. Roychowdhury, 2004, “Independent Component Analysis Based Nonparametric Density Estimation,” IEEE Transaction on Neural Networks, Vol. 15, pp. 55-65.

Collins, R. T., A. J. Lipton and T. Kanade, 2000, “A system for video surveillance and monitoring.” VSAM final report. Carnegie Mellon University, Technical Report, CMU-RI-TR-00-12, 2000.

Comon, P., 1994, “Independent component analysis: a new concept?,’’ Signal Processing, Vol. 36, pp. 287-314.

Cutler, R and L. R. Davis, 2000, “Real-time periodic motion detection, analysis, and applications.” IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 22, pp. 781-796.
Eberhart, R. C. and J. Kennedy, 1995, “A new optimizer using particle swarm theory,’’ Proceeding of Sixth International Symposium on Micro Machine and Human Science, IEEE Service Center, Nagoya, Japan, pp. 39-43.

Haritaoglu I., D. Harwood and L. Davis, 2000, “Real-time surveillance of people and their activities,” IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 22, pp. 809-830.

Hawthorne J., 2000, “Electro-optics technology test flat-panel displays,” Automated Testing, May, pp. 271-276.

Hu W., T. Tan, L. Wang and S. Maybank, 2004, “A survey on visual surveillance of object motion and behaviors,” IEEE Trans. on Systems, Man and Cybernetics, Part C: Applications and Reviews, Vol. 34, pp. 334-352.

Hung C. I, P. L. Lee, Y. T. Wu, L. F. Chen, T. C. Yeh, J. C. Hsieh, 2005, “Recognition of Motor Imagery Electroencephalography Using Independent Component Analysis and Machine Classifiers.” Annals of Biomedical Engineering, Vol. 33, pp. 1053-1070.

Hyvärinen, A. P. Hoyer and E. Oja, 1998, “Sparse code shrinkage for image denoising,’’ Proceedings of IEEE International Joint Conference on Neural Networks, Anchorage, Alaska, pp. 859-864.

Hyvärinen, A., 1998, “New approximations of differential entropy for independent component analysis and projection pursuit,’’ In Advances in Neural Information Processing Systems, pp. 273-279.

Hyvärinen, A., 1999a, “The fixed-point algorithm and maximum likelihood estimation for independent component analysis,’’ Neural Processing Letters, Vol. 10, pp. 1-5.

Hyvärinen, A., 1999b, “Sparse code shrinkage: denoising of nongaussian data by maximum likelihood estimation,’’ Neural Computation, Vol. 11, pp. 1739-1768.

Hyvärinen, A., 1999c, “Fast and robust fixed-point algorithms for independent component analysis,’’ IEEE Trans. on Neural Networks, Vol. 10, pp. 626-634.

Hyvärinen, A. and E. Oja, 2000, “Independent component analysis: algorithms and applications,’’ Neural Networks, Vol. 13, pp. 411-430.

Hyvärinen, A., P. Hoyer and E. Oja, 2001a, “Image denoising by Sparse code shrinkage,’’ In: Intelligent Signal Processing (Ed. S. Haykin and B. Kosko), IEEE Press, New York, pp. 554-568.

Ikeda, S. and K. Toyama, 2000, “Independent component analysis for noisy data-MEG data analysis,’’ Neural Networks, Vol. 13, pp. 1063-1074.

Jenssen, R. and T. Eltoft, 2003, “Independent component analysis for texture segmentation,’’ Pattern Recognition, Vol. 36, pp. 2301-2315.

Kao Y. H., W. Y. Guo, Y. T. Wu, K. C. Liu,W. Y. Chai, C. Y. Lin, Y. S. Hwang, A. K. Liou, H. C. Cheng, T. C. Yeh, J. C. Hsieh, M. M. H. Teng, 2003, “Hemodynamic Segmentation of MR Brain Perfusion Images Using Independent Component, Bayesian Estimation and thresholding,” Magnetic Resonance in Medicine, Vol. 49, pp.885-894.

Kido. T., 1992, “In-process inspection technique for active-matrix LCD panels,” International Test Conference, Maryland, USA, pp. 795-799.

Kido. T., 1993, “In-process functional inspection technique for TFT-LCD array,” Journal of SID, Vol. 1, pp. 429-435.

Kido. T., 1995, “Optical charge-sensing method for testing and characterizing thin-film transistor array,” IEEE Journal of Selected Topics in Quantum Electronics, Vol. 1, pp. 993-1001.

Kim, J. H., S. Ahn., J. W. Jeon, and J. E. Byun, 2001, “A high-speed high-resolution vision system for the inspection of TFT LCD,” IEEE International Symposium on Industrial Electronics, Vol. 1, pp. 12-16.

Lipton A, H. Fujiyoshi and R. Patil, 1998, “Moving target classification and tracking from real-time video.” In: Proc. IEEE Workshop on Applications of Computer Vision, Princeton, NJ, 8-14.

Meyer D, J. Denzler and H. Niemann, 1997, “Model based extraction of articulated objects in image sequences for gait analysis.” In: Proc. IEEE International Conference on Image Processing, Santa Barbara, California pp. 78-81.

Roehrig, H., E. A. Krupinski, A. S. Chawla, J. Fan, K.Gandh, T. Furukawa, and M. Ohashii, 2003, “Noise of LCD display systems,” International Congress Series, Vol. 1256, pp. 162-168.

Roehrig, H., E. A. Krupinski, J. Fan, and K.Gandhi, 2004, “Physical and psychophysical evaluation of LCD noise,” International Congress Series, Vol. 1268, pp. 341-346.

Salman, A., I. Ahmad and S. Al-Madani, 2002, “Particle swarm optimization for task assignment problem,”Microprocessors and Microsystems, Vol. 26, pp. 363-371.

Shi, Y. and R. C. Eberhart, 1998, “A modified particle swarm optimizer,’’ ” In: Proc. IEEE International Conference on Evolutionary Computation, Anchorage, Alaska, pp. 69-73.

Stauffer C and W. A. Grimson, 1999, “Background mixture models for real-time tracking.” In: Proc. IEEE Conference on Computer Vision and Pattern Recognition, Fort Collins, Colorado, pp. 246-252.

Tsai D. M. and C. Y. Hung, 2005, “Automatic Defect Inspection of Patterned TFT-LCD Panels Using 1-D Fourier Reconstruction and Wavelet Decomposition,” International Journal of Production Research, Vol 43, pp. 4589-4607