book

Intelligent Speech Signal Processing

by Nilanjan Dey

March 2019

Intermediate to advanced

209 pages

6h 20m

English

Academic Press

Read now

Unlock full access

Abstract
Abstract2.1 Introduction2.2 Related Work2.3 Various Architecture of ASR2.4 Convolutional Neural Networks2.5 CNN-Based End-to-End Approach2.6 Experiments and Their Results2.7 Conclusion
Abstract3.1 Introduction3.2 Microchip dsPIC33 Digital Signal Controller3.3 High Pass Filter3.4 Fast Fourier Transform3.5 Channel Energy Computation3.6 Channel SNR Computation3.7 VAD Decision3.8 VAD Hangover3.9 Computation of Scaling Factor3.10 Scaling of Frequency Channels3.11 Inverse Fourier Transform3.12 Application Programming Interface3.13 Resource Requirements3.14 Microchip PIC Programmer3.15 Audio Components3.16 VAD and Background Noise Reduction Techniques3.17 Results and Discussion3.18 Conclusion and Discussion

Abstract4.1 Introduction4.2 Detection of Conflicting Classes4.3 Complementary Models for Classification4.4 Proposed Cascade of Classifiers4.5 Audio-Visual Databases4.6 Experimental Results4.7 Conclusions
Abstract5.1 Introduction5.2 Spoken Language Identification5.3 Cues for Spoken Language Identification5.4 Stages in Spoken Language Identification5.5 Deep Learning5.6 Artificial and Deep Neural Network5.7 Comparison of Spoken LID System Implementations with Deep Learning Techniques5.8 Discussion5.9 Conclusion
Abstract6.1 Introduction6.2 Conceptual Design of the System6.3 System Implementation6.4 Significance/Contribution6.5 Conclusion
Abstract7.1 Introduction7.2 Extractive Summarization7.3 Abstractive Summarization7.4 Need for Speech Summarization7.5 Issues in the Summarization of a Spoken Document7.6 Tamil Language7.7 System Design for Summarization of Speech Data in Tamil Language7.8 Evaluation Metrics7.9 Speech Corpora for Tamil Language7.10 Conclusion
Abstract8.1 Introduction8.2 Data Collection and Processing8.3 Research Methodology8.4 Numerical Experiments and Results8.5 Conclusion
Abstract9.1 Wavelet Packet Decomposition9.2 Empirical Mode Decomposition9.3 Variational Mode Decomposition9.4 Synchrosqueezing Wavelet Transform: EMD Like a Tool9.5 Applications of the Decomposition Technique9.6 Conclusion
Abstract10.1 Introduction10.2 Literature Survey10.3 AWS IoT10.4 Amazon Voice Service (AVS)10.5 AWS Lambda10.6 Message Queuing Telemetry Transport (MQTT)10.7 Proposed Architecture10.8 Conclusion

Content preview from Intelligent Speech Signal Processing

Chapter 2

End-to-End Acoustic Modeling Using Convolutional Neural Networks

Vishal Passricha; Rajesh Kumar Aggarwal National Institute of Technology Kurukshetra, Kurukshetra, India

Abstract

State-of-the-art automatic speech recognition (ASR) systems map speech into its corresponding text. Conventional ASR systems model the speech signal into phones in two steps; feature extraction and classifier training. Traditional ASR systems have been replaced by deep neural network (DNN)-based systems. Today, end-to-end ASRs are gaining in popularity due to simplified model-building processes and the ability to directly map speech into text without predefined alignments. These models are based on data-driven learning methods and competition with complicated ...