ICA and Binary-Mask-Based Blind Source Separation with Small Directional Microphones

• Published in: Independent Component Analysis and Blind Signal Separation pp. 649 - 657
• Main Authors: Mori, Yoshimitsu, Saruwatari, Hiroshi, Takatani, Tomoya, Shikano, Kiyohiro, Hiekata, Takashi, Morita, Takashi
• Format: Book Chapter Conference Proceeding
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 01.01.2006; Springer
• Series: Lecture Notes in Computer Science
• Subjects: Applied sciences; Binary Mask; Blind Source Separation; Exact sciences and technology; Independent Component Analysis; Information, signal and communications theory; Signal processing; Sound Source; Telecommunications and information theory; Signal mixing; Microphone; Source separation; Mask; Sound quality; Blind source separation; Software component; Sound source; Modeling; Blind separation; Separation method; Signal processing; Signal sources; Principal component analysis; Input output model
• ISBN: 3540326308; 9783540326304
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/11679363_81

A new two-stage blind source separation (BSS) for convolutive mixtures of speech is proposed, in which a Single-Input Multiple-Output (SIMO)-model-based ICA and binary mask processing are combined. SIMO-model-based ICA can separate the mixed signals, not into monaural source signals but into SIMO-model-based signals from independent sources as they are at the microphones. Thus, the separated signals of SIMO-model-based ICA can maintain the spatial qualities of each sound source. Owing to the attractive property, binary mask processing can be applied to efficiently remove the residual interference components after SIMO-model-based ICA. The experimental results using small directional microphone array reveal that the separation performance can be considerably improved by using the proposed method in comparison to the conventional source separation methods.