A Modified Fast Independent Component Analysis and Its Application to Image Separation
一种改进的快速独立分量分析算法及其在图象分离中的应用

Independent Component Analysis (ICA) is a new development of signal processing. As an effective approach to the separation of blind signal, Independent Component Analysis has attracted broad attention and has been successfully used in many fields. The fundamental, discrimination condition and practical algorithm of Independent Component Analysis are discussed. Then, a fast Independent Component Analysis algorithm (FastICA) is introduced, and it is known that the time-consuming course is computing Jacobian Matrix. Reducing the time of Jacobian Matrix will improve the performance of FastICA algorithm. So a modified FastICA (M-FastICA) algorithm is developed. By modifying kernel iterate course, several iterations of FastICA are merged into one iteration of M-FastICA, then M-FastICA algorithm only need to compute Jacobian Matrix once time and achieves the correspondent effect of FastICA. So the convergence of ICA will be accelerated. Finally, M-FastICA is applied to image separation. The experiment images are mixed with a random matrix. Independent Component Analysis can separate the mixed images and obtain the approximate of source images. The experiment results show that the iterations of serial modified algorithm reduces 9 percent, and the iterations of parallel modified algorithm reduces 27 percent with the correspondent separation performance.