Blind Source Separation and Human Speech Extraction for Three Sound Sources Using Silent Interval
Kana Higuchi, Chiharu Okuma, Takaaki Ishibashi · 2017
In this paper, we propose a blind source separation method for multiple sound source signals. We have proposed an estimation method for source signals without noise from observed mixture signals. The method estimates rotation angle of a distribution of observed mixture signals. The distribution is rotated based on the estimated angle. Our method can estimate the source signals from two mixture signals. However, in case of mixture signals, the source signals cannot be estimated by one processing. Therefore, we apply processing multiple times to observed mixture signals. In this way, we propose an estimation method for the original source.