Auxiliary function-based algorithm for blind extraction of a moving speaker

Abstract In this paper, we propose a novel algorithm for blind source extraction (BSE) of moving acoustic recorded by multiple microphones. The is based on independent vector (IVE) where the contrast function optimized using auxiliary function-based technique and recently proposed constant separating (CSV) mixing model assumed. CSV allows movements extracted within analyzed batch recordings. We provide practical explanation how works when extracting source. Then, experimentally verified task speaker. compared with state-of-the-art methods an adaptive BSE which processes data in sequential manner. results confirm that can extract speaker better than conventional it achieves improved accuracy method.