For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
A Speech Enhancement Algorithm Based on Blind Signal Cancelation in Diffuse Noise Environments
Jaesik HWANG Jaepil SEO Ji-Won CHO Hyung-Min PARK
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 2016/01/01
Online ISSN: 1745-1337
Type of Manuscript: LETTER
Category: Speech and Hearing
speech enhancement, Wiener filtering, beamforming, blind signal extraction,
Full Text: PDF(245.5KB)>>
This letter describes a speech enhancement algorithm for stereo signals corrupted by diffuse noise. It estimates the noise signal and also a beamformed target signal based on blind target signal cancelation derived from sparsity minimization. Enhanced target speech is obtained by Wiener filtering using both the signals. Experimental results demonstrate the effectiveness of the proposed method.