Abstract An improved fast independent component analysis (ICA) feature extraction algorithm is proposed based on one dimension search and distance function. Aiming at the problem that the selection of the initial value influences the convergence in fast ICA algorithm, the improved algorithm ensures the good convergence result by using one dimension search. Meanwhile, a rule based on distance function is designed to select the optimal features for the recognition according to the characteristics of the infrared image. Thus, the problem of the recognition rate and the robustness decreasing with the increasing number of training image samples is resolved. Compared with the traditional methods, the experimental results of real infrared images show that the proposed algorithm reaches a lower error classification rate with fewer infrared object features, and the error classification rate of the proposed method is robust in different kinds of classes.
[1] Bayik T M. Automatic Target Recognition in Infrared Imagery. Ankara, Turkey: Middle East Technical University Press, 2004 [2] Hyvarinen A, Karhunen J, Oja E. Independent Component Analysis. New York, USA: Wiley, 2001 [3] Jolliffe I T. Principal Component Analysis. 2nd Edition. Alexandria, USA: American Statistical Association, 2003 [4] Hyvrinen A. Blind Source Separation by Nonstationarity of Variance: A Cumulant-Based Approach. IEEE Trans on Neural Networks, 2001, 12(6):1471-1474 [5] Hyvrinen A. Fast and Robust Fixed-Point Algorithms for Independent Component Analysis. IEEE Trans on Neural Networks, 1999,
10(3): 626-634 [6] Shyu K K, Lee M H, Wu Y T, et al. Implementation of Pipelined Fast ICA on FPGA for Real-Time Blind Source Separation. IEEE Trans on Neural Networks, 2008, 19(6): 958-970 [7] Hao Shen, Huper K. Generalised Fastica for Independent Subspace Analysis // Proc of the IEEE International Conference on Acoustics, Speech and Signal Processing. Honolulu, USA, 2007, Ⅳ: 1409-1412 [8] Wang Mingxiang, Mo Yulong. A New Method for the Detection of Moving Target Based on Fast Independent Component Analysis. Computer Engineering, 2004, 30(3): 58-60 (in Chinese) (王明祥,莫玉龙.一种基于Fast ICA的运动目标检测新方法.计算机工程, 2004, 30(3): 58-60) [9] Dun Wenjie, Mu Zhichun. An ICA-Based Ear Recognition Method through Nonlinear Adaptive Feature Fusion. Journal of Computer-Aided Design Computer Graphics, 2009, 21(3): 382-388 (in Chinese) (敦文杰,穆志纯.基于ICA的非线性自适应特征融合的人耳识别.计算机辅助设计与图形学学报, 2009, 21(3): 382-388) [10]Zou Muchun. Face Recognition Based on Fast ICA and RBF Neural Networks // Proc of the International Symposium on Information Science and Engineering. Shanghai, China, 2008, Ⅰ: 588-592 [11]Li Ming, Wu Fuwen, Liu Xueyan. Face Recognition Based on WT, Fast ICA and RBF Neural Network // Proc of the 3rd International Conference on Natural Computation. Qingdao, China, 2007, Ⅱ: 3-7 [12]Chen Baolin. Optimization Theory and Algorithms. 2nd Edition. Beijing, China: Tsinghua University Press, 2005 (in Chinese) (陈宝林.最优化理论与算法.第2版.北京:清华大学出版社, 2005)
2010, Vol. 23 
