Handbook of Neural Networks for Speech Processing

Citation

Rigelsford, J. (2003), "Handbook of Neural Networks for Speech Processing", Sensor Review, Vol. 23 No. 4. https://doi.org/10.1108/sr.2003.08723dae.003

Handbook of Neural Networks for Speech Processing

Handbook of Neural Networks for Speech Processing

S. KatagiriArtech House2000522 pp.ISBN 0-89006-954-9£39.00

Keywords: Neural networks, Speech recognition

This book provides the fundamentals and practical applications of neural network based speech processing. Its comprehensive coverage is easy to read and is complemented by the outstanding illustrations used throughout the text. The book is suitable for students, practising engineers and researchers.

The "Handbook of Neural Networks for Speech Processing" comprises 13 chapters divided into three parts. Part I, Fundamentals, introduces the basics of speech processing and artificial neural networks. After an introduction, chapter 2 addresses the Speech Signal and its Production Model. Chapters 3 and 4 discuss Speech Recognition, and Speech Coding, respectively. Topics addressed include: acoustic feature extraction; language modelling; and speech coding over noisy channels.

Part II, comprises five chapters presenting Current Issues in Speech Recognition. Discriminative Prototype-Based Methods for Speech Recognition are presented in chapter 5, while chapter 6 discusses Recurrent Neural Networks for Speech Recognition. The following chapter addresses Time-Delay Neural Networks (TDNN) and neural network, hidden Markov model (NN/HMM) Hybrids: A Family of Connectionist Continuous-Speech Recognition Systems. Chapters 8 and 9 discuss Probability-Oriented Neural Networks and Hybrid Connectionist/ Stochastic Networks, and Minimum Classification Error Networks, respectively.

The final part of the book addresses Current Issues in Speech Signal Processing. Networks for Speaker Recognition are presented in chapter 10, while chapter 11 discusses Neural Networks for Voice Conversion. Quantization performance of neural networks, and nonlinear prediction speech coding, are amongst the topics addressed in chapter 12, Neural Networks for Speech Coding. Chapter 13, Networks for Speech Enhancement, discusses topics including: neural time-domain filtering methods, neural transform-domain methods, and state-dependent model switching methods.

Overall, the "Handbook of Neural Networks for Speech Processing" is an excellent reference text for anyone interested in the rapidly growing, multidisciplinary area of research.

Jon Rigelsford