Information Technology Journal 1812-5638 1812-5646 Asian Network for Scientific Information 10.3923/itj.2012.456.460 Wen Xinling Miao Manxiang 4 2012 11 4 This paper studied the Independent Component Analysis (ICA) algorithm, mainly introduces the ICA model and design an algorithm based on the Fast ICA method, which is the most widely used in blind source separation at present. A kind of Fast ICA algorithm is researched in this paper and through simulation of three mixed images, we prove this algorithm’s accuracy in separating blind sources signal, as well as obtain precise result in image blind separation. From the simulation diagram results we can see, that Fast ICA algorithm used in images blind separation can get better separation effect.

]]> 1 Hyvarinen, A., 1999 Neural Comput. Surv., 2 94 128 2 Comon, P., 1994 Signal Process., 36 287 314 3 Cuiru, W., D. Juan and Z. Xinlin, 2006 Chin. J. Sci. Instrum., 27 785 786 4 Yuen, P.C. and J.H. Lai, 2001 Pattern Recognit., 34 545 553 5 Guo, S. and Y.L. Sun, 2004 Electron. Meas. Technol., 27 12 14 6 Yang, F. and B. Hong, 2006 The Principle and Application of ICA. 7 Yang, B., 2007 Research and application of BBS algorithm based on fast ICA. 8 Hyvarinen, A., J. Karhunen and E. Oja, 2001 Independent Component Analysis. 9 Tian, Q., W. Zheng and D. Sun, 2009 Skills Maintenance, 16 90 91 10 Hyvarinen, A. and J. Karhunen, 2007 ICA, Zhou Zongtan. pp: 191-192 pp: 191-192 11 Amari, S., T. Chen and A. Cichocki, 1997 Non-Holonomic Constraints in Learning Blind Source Separation. 1997 pp: 633-636 pp: 633-636