ISS2: An Extension of Iterative Source Steering Algorithm for Majorization-Minimization-Based Independent Vector Analysis
Abstract
A majorization-minimization (MM) algorithm for independent vector analysis optimizes a separation matrix by minimizing a surrogate function of the form , where is the number of sensors and positive definite matrices are constructed in each MM iteration. For , no algorithm has been found to obtain a global minimum of . Instead, block coordinate descent (BCD) methods with closed-form update formulas have been developed for minimizing and shown to be effective. One such BCD is called iterative projection (IP) that updates one or two rows of in each iteration. Another BCD is called iterative source steering (ISS) that updates one column of the mixing matrix in each iteration. Although the time complexity per iteration of ISS is times smaller than that of IP, the conventional ISS converges slower than the current fastest IP (called ) that updates two rows of in each iteration. We here extend this ISS to that can update two columns of in each iteration while maintaining its small time complexity. To this end, we provide a unified way for developing new ISS type methods from which as well as the conventional ISS can be immediately obtained in a systematic manner. Numerical experiments to separate reverberant speech mixtures show that our converges in fewer MM iterations than the conventional ISS, and is comparable to .
Keywords
Cite
@article{arxiv.2202.00875,
title = {ISS2: An Extension of Iterative Source Steering Algorithm for Majorization-Minimization-Based Independent Vector Analysis},
author = {Rintaro Ikeshita and Tomohiro Nakatani},
journal= {arXiv preprint arXiv:2202.00875},
year = {2022}
}
Comments
Accepted for publication in the 30th European Signal Processing Conference (EUSIPCO 2022)