This revised scholarly work on voice recognition technology outlines cutting-edge research in this exciting area of computer science. The book begins with a readable historical introduction to speech synthesis, speech recognition, and speaker classification. (According to the authors, Alexander Graham Bell was actually working on the problem of speech synthesis when he invented the telephone.)

Once the authors move on to their specialty, this text gets increasingly advanced, discussing the multitude of formulas and algorithms used in speech recognition. The writers clearly explain two important areas of recognizing speech, Hidden Markov Models (HMMs) and neural nets as applied to speech.

The latter part of the book presents the authors' own research in the field, which involves using the technology to identify characteristics of voices, enabling law enforcement to catch criminals who commit crimes over phone lines. (One of their principal advances, "rehumanizing filter techniques" enables a voice to be identified in even noisy environments.) Though the authors certainly would be the first to admit that voice recognition has a long way to go before it becomes commonplace, they see a strong future for this technology in law and security.


From Book News, Inc.
Klevans and Rodman (North Carolina State U.) identify the diverse audience with interests in computer voice-recognition systems: researchers (in computer, speech, and human-factors); security (banks, corporations, the military, homeowners); and professionals in justice, law enforcement, education, and training. The linguist and computer scientist duo examine the major implicated tasks of feature extraction and pattern recognition, review its history since voice-print analysis of the 1960s, compare human/machine voice recognition, consider design tradeoffs, present their multigrammar segregating system for context- free voice recognition, share experimental results, and make recommendations regarding specific applications and a future research agenda. Book News, Inc.®, Portland, OR


Book Description
Here's a scientific look at computer-generated speech verification and identification -- its underlying technology, practical applications, and future direction. This book provides a solid background in voice recognition technology and enables informed decision-making on how to select the most appropriate voice recognition-based software for a particular company or organization. It provides clear explanations of mathematical concepts, as well as a full-chapter presentation of the successful new Multi-Granular Segregating System for accurate, context-free speech identification.

Page after page of actual case studies and experimental results supported by clear, easy-to-follow charts and graphs explain the technology behind voice recognition systems, describe the advantages and disadvantages of the available systems, and show the benefits and limitations of forensic voice recognition technology.


About the Author
Richard Rodman is an associate professor of computer science at North Carolina State University. The author of two books and more than a dozen professional papers, he holds a Ph.D. in linguistics from UCLA.

FROM THE PUBLISHER

Here's a scientific look at computer-generated speech verification and identification — its underlying technology, practical applications, and future direction. This book provides you with a solid background in voice recognition technology and helps you make informed decisions on which voice recognition-based software to use in your company or organization. It is unique in its clear explanations of mathematical concepts, as well as its full-chapter presentation of the successful new Multi-Granular Segregating System for accurate, context-free speech identification. Page after page of actual case studies and experimental results supported by clear, easy-to-follow charts and graphs help... ￯﾿ﾽ Speech scientists get up to speed in voice recognition technology and conduct future research ￯﾿ﾽ Security specialists and managers choose the speech identification-based security system best suited for their needs ￯﾿ﾽ Law enforcement and justice officials better understand the benefits and limitations of forensic voice recognition technology

FROM THE CRITICS

Booknews

Klevans and Rodman (North Carolina State U.) identify the diverse audience with interests in computer voice-recognition systems: researchers (in computer, speech, and human-factors); security (banks, corporations, the military, homeowners); and professionals in justice, law enforcement, education, and training. The linguist and computer scientist duo examine the major implicated tasks of feature extraction and pattern recognition, review its history since voice-print analysis of the 1960s, compare human/machine voice recognition, consider design tradeoffs, present their multigrammar segregating system for context- free voice recognition, share experimental results, and make recommendations regarding specific applications and a future research agenda. Annotation c. by Book News, Inc., Portland, Or.

ACCREDITATION