Speech Signal Recovery Using Block Sparse Bayesian Learning

• Published in: Arabian journal for science and engineering (2011) Vol. 45; no. 3; pp. 1567 - 1579
• Main Authors: Ahmed, Irfan, Khan, Aftab, Ahmad, Nasir, NasruMinallah, Ali, Hazrat
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.03.2020; Springer Nature B.V
• Subjects: Audio data; Bayesian analysis; Datasets; Engineering; Humanities and Social Sciences; Machine learning; multidisciplinary; Recovery; Research Article - Electrical Engineering; Science; Signal quality; Signal reconstruction; Similarity; Speech; SSIM; Signal recovery; BSBL; Compressed sensing; Wavelet denoising
• ISSN: 2193-567X; 1319-8025; 2191-4281
• DOI: 10.1007/s13369-019-04080-6

Compressed sensing is based on the recovery of original signal from the low-quality and incomplete samples. Recently, ℓ 1 -norm is used for the estimation of signal elements from the underdetermined set of equations. In this paper, we propose a technique for speech signal recovery called block sparse Bayesian learning. The proposed technique is applied over the random set of speech samples and acquired better performance as compared to ℓ 1 -based recovery. Apart from the proposed recovery technique, this work is also intended to develop a trained and efficient sampling matrix through offline training. In this work, we apply structural similarity index as a metric to compare the performance of the proposed technique with an existing ℓ 1 based recovery. Sparse Bayesian learning and ℓ 1 -norm recovery are applied over the selected audio files from the datasets. The dataset consists of speech signals from three different languages: Urdu, Pashto and English. Structural similarity between the recovered and original speech signals is used as a metric to compare the performance of BSBL with ℓ 1 -norm minimization. The comparison based on structural similarity index shows the effectiveness of the proposed technique.