A new efficient two-channel fast transversal adaptive filtering algorithm for blind speech enhancement and acoustic noise reduction. (January 2019)

• Record Type: Journal Article
• Authors: Henni, Rahima; Djendi, Mohamed; Djebari, Mustapha
• Is Part Of: Computers & electrical engineering. Volume 73(2019)
• Page Start: 349
• Page End: 368
• Publication Date: 2019-01
• Subjects: Adaptive algorithm -- Speech enhancement -- BSS -- Two-channel algorithm -- MSE -- SNR; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2018.12.009
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: We propose a two-channel backward BSS of the simplified FTF (TCSFTF) algorithm for speech enhancement. We propose to use this new two-channel algorithm combined with the backward BSS structure to enhance the speech signal blindly. Intensive experiments have shown the good performance and the robustness of this proposed algorithm. A comparison with other competitive and state-of-the-art algorithms have shown the superiority of the proposed algorithm. Abstract: This paper addresses the problem of speech enhancement and acoustic noise reduction by blind structures. Recently, the backward blind source separation (BBSS) structure has shown efficiency in cancelling the acoustic noise and improving corrupted speech signals form very noisy observations without any a priori information of source signals. In this paper, we propose a new algorithm based on the combination between the BBSS structure and the simplified fast transversal filter (SFTF) algorithm. The proposed two-channel simplified fast transversal filter (TCSFTF) algorithm succeeded an important blind improvement of steady state and convergence speed performances in diverse noisy situations when only the noisy signals are known. The performances of the new TCSFTF algorithm are compared with four state-of-the-art algorithms in different noisy conditions. This comparison is evaluated in terms of cepstral distance (CD), system mismatch (SM), segmental signal to noise ratio (SegSNR), and segmental mean square error (SegMSE) criteria. Graphicalabstract : Image, graphical abstract