The Stationary Bionic Wavelet Transform and its Applications for ECG and Speech Processing (Signals and Communication Technology)

✍ Scribed by Talbi Mourad

• Publisher: Springer
• Year: 2022
• Tongue: English
• Leaves: 95
• Category: Library

✦ Synopsis

This book first details a proposed Stationary Bionic Wavelet Transform (SBWT) for use in speech processing. The author then details the proposed techniques based on SBWT. These techniques are relevant to speech enhancement, speech recognition, and ECG de-noising. The techniques are then evaluated by comparing them to a number of methods existing in literature. For evaluating the proposed techniques, results are applied to different speech and ECG signals and their performances are justified from the results obtained from using objective criterion such as SNR, SSNR, PSNR, PESQ , MAE, MSE and more.

✦ Table of Contents

Book Summary
Introduction
Contents
About the Author
Acronyms
Chapter 1: Speech Enhancement Based on Stationary Bionic Wavelet Transform and Maximum A Posterior Estimator of Magnitude-Squared Spectrum
1.1 Introduction
1.2 The Proposed Technique
1.2.1 Background
1.2.1.1 Wavelet Analysis [21, 22]
1.2.1.2 The Bionic Wavelet Transform
1.2.1.3 The Stationary Bionic Wavelet Transform (SBWT)
The Stationary Wavelet Transform (SWT)
Perfect Reconstruction of SBWT
1.3 Application of the Maximum A Posterior Estimator of Magnitude-Squared Spectrum in SBWT Domain
1.4 The Evaluation Metrics
1.4.1 Signal-to-Noise Ratio
1.4.2 Segmental Signal-to-Noise Ratio
1.4.3 Itakura–Saito Distance
1.4.4 Perceptual Evaluation of Speech Quality (PESQ)
1.5 Results and Discussions
1.6 Conclusion
References
Chapter 2: ECG Denoising Based on 1-D Double-Density Complex DWT and SBWT
2.1 Introduction
2.2 Materials
2.2.1 The BWT Optimization for ECG Analysis
2.2.2 1-D Double-Density Complex DWT
2.2.3 Denoising Technique Based on Wavelets and Hidden Markov Models
2.2.4 The Denoising Approach Based on Non-local Means
2.2.5 The ECG Denoising Approach Based on BWT and FWT_TI [49]
2.2.6 The Proposed ECG Denoising Approach [29]
2.3 Results and Discussion
2.4 Conclusion
References
Chapter 3: Speech Enhancement Based on SBWT and MMSE Estimate of Spectral Amplitude
3.1 Introduction
3.2 The MMSE Estimate of Spectral Amplitude
3.2.1 Signal Model
3.3 The Proposed Speech Enhancement Approach [19]
3.4 Minimum Mean Square Error (MMSE) Estimate of Spectral Amplitude in the SBWT Domain
3.5 Unsupervised Speech Denoising Via Perceptually Motivated Robust Principal Component Analysis [23]
3.6 The Speech Enhancement Technique Based on MSS–SMPO [25]
3.7 Results and Discussion
3.8 Conclusion
References
Chapter 4: Arabic Speech Recognition by Stationary Bionic Wavelet Transform and MFCC Using a Multi-layer Perceptron for Voice Control
4.1 Introduction
4.2 The Feature Extraction
4.2.1 MFCC Extraction
4.3 Pre-emphasis
4.4 Frame Blocking and Windowing
4.5 DFT Spectrum
4.6 Mel Spectrum
4.7 Discrete Cosine Transform (DCT)
4.8 Dynamic MFCC Features
4.9 Classifiers
4.10 The Proposed Speech Recognition Technique [12]
4.11 Experiments and Results
4.12 Conclusion
References
Index