Blind separation of multidimensional mixed signals based on fast independent component analysis
Wu Zheng-Mao, Yunlian Sun · 2005
A new fast algorithm of independent component analysis(FastICA) applied to blind signal separation is introduced.In the method, negentropy of information theory is used as the objective function which estimates statistical independence between output components . A fast iterative algorithm is derived by optimizing the objective function .The method does not need to calculate the higher order statistics of output components and converges fast. The simulation experiment on the separation of linear mixed time signals and image signals are done by FastICA algorithm. The efficiency of the FastICA algorithm is verified by the simulation experiment results.