Postprocessing With Wiener Filtering Technique for Reducing Residual Crosstalk in Blind Source Separation

Abstract

In this letter, a postprocessing method using the Wiener filtering technique is proposed to reduce the residual crosstalk in frequency-domain independent component analysis (FDICA). By the proposed method, the target signal components remain with little attenuation while the interference components are suppressed. The experimental results show that the proposed method reduces the residual crosstalk and slightly improves the separation performance over the postprocessing using the normalized least-mean-square (NLMS) algorithm by about 1-2 dB with much less computation