A New ICA Algorithm Based on CDF Estimation

Author

Shangzhi, Xu ; Aihuang, Guo

Author_Institution

Dept.of Inf. & Commun., TongJi Univ., Shanghai, China

fYear

2010

fDate

25-27 June 2010

Firstpage

4908

Lastpage

4911

Abstract

The technique of Independent component analysis can be applied in image processing, communication processing, data mining and many others. By research on score function of informax algorithm and the performance of Ext-informax algorithm while the sources include impulse signals and Gaussian signals. The results show these algorithms are not stably in these instances. A new score function according to cumulative probability distribution function of signal is proposed. All simulations show better performance than Ext-Informax algorithm under above instances.

Keywords

Gaussian distribution; blind source separation; independent component analysis; CDF estimation; Ext-informax algorithm; Gaussian signal; ICA algorithm; cumulative probability distribution function; independent component analysis; Algorithm design and analysis; Educational institutions; Estimation; Image processing; Independent component analysis; Signal processing; Signal processing algorithms; Informax algorithm; cumulative distribution function; gaussian signal; independent component analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Control Engineering (ICECE), 2010 International Conference on

Conference_Location

Wuhan

Print_ISBN

978-1-4244-6880-5

Type

conf

DOI

10.1109/iCECE.2010.1187

Filename

5630008