Robust Adaptation to Non-Native Accents in Automatic Speech Recognition

Speech recognition technology is being increasingly employed in human-machine interfaces. A remaining problem however is the robustness of this technology to non-native accents, which still cause considerable difficulties for current systems.
In this book, methods to overcome this problem are described. A speaker adaptation algorithm that is capable of adapting to the current speaker with just a few words of speaker-specific data based on the MLLR principle is developed and combined with confidence measures that focus on phone durations as well as on acoustic features. Furthermore, a specific pronunciation modelling technique that allows the automatic derivation of non-native pronunciations without using non-native data is described and combined with the previous techniques to produce a robust adaptation to non-native accents in an automatic speech recognition system.

List of contents

ASR:AnOverview.- Pre-processing of the Speech Data.- Stochastic Modelling of Speech.- Knowledge Bases of an ASR System.- Speaker Adaptation.- Confidence Measures.- Pronunciation Adaptation.- Future Work.- Summary.- Databases and Experimental Settings.- MLLR Results.- Phoneme Inventory.

Summary

Speech recognition technology is being increasingly employed in human-machine interfaces. A remaining problem however is the robustness of this technology to non-native accents, which still cause considerable difficulties for current systems.
In this book, methods to overcome this problem are described. A speaker adaptation algorithm that is capable of adapting to the current speaker with just a few words of speaker-specific data based on the MLLR principle is developed and combined with confidence measures that focus on phone durations as well as on acoustic features. Furthermore, a specific pronunciation modelling technique that allows the automatic derivation of non-native pronunciations without using non-native data is described and combined with the previous techniques to produce a robust adaptation to non-native accents in an automatic speech recognition system.

Product details

• Authors: S. Goronzy, Silke Goronzy
• Publisher: Springer, Berlin
• Languages: English
• Product format: Paperback / Softback
• Released: 19.12.2002
• EAN: 9783540003250
• ISBN: 978-3-540-00325-0
• No. of pages: 146
• Weight: 266 g
• Illustrations: XI, 146 p.
• Series: Lecture Notes in Computer Science; Lecture Notes in Artificial Intelligence
• Subjects: Natural sciences, medicine, IT, technology > IT, data processing > IT; Elektronik, B, Artificial Intelligence, Mensch-Computer-Interaktion, Theoretische Informatik, computer science, Digitale Signalverarbeitung (DSP), Signal, Image and Speech Processing, Signal Processing, Mathematical theory of computation, Speech processing systems, Imaging systems & technology, Image processing, Mathematical logic, Mathematical Logic and Formal Languages, User interface design & usability, User Interfaces and Human Computer Interaction, User interfaces (Computer systems), Signal, Speech and Image Processing, Formal Languages and Automata Theory