In the very recent past, automatic speech recognition (ASR) systems have attained very acceptable performance when used with speech captured with a head-mounted or close-talking microphone (CTM). But the performance of conventional ASR systems degrades dramatically as soon as the microphone is moved away from the mouth of the speaker; this degradation is due to a broad variety of effects that are not found in CTM speech, including background noise, overlapping speech from other speakers, and reverberation. This makes conventional ASR systems all but useless in many scenarios where they would otherwise be most helpful; i.e., in situations such as giving spoken directions to a humanoid robot, asking a car navigation system for directions, or in a meeting scenario where amicrophone cannot be mounted on a speaker's head. While conventional ASR systems perform miserably for speech captured with far-field sensors, there are a number of techniques developed in other areas of signal processing that can mitigate the deleterious effects of noise and reverberation, as well as separating speech from overlapping speakers. In this book, these techniques will be made accessible and comprehensible to a broad audience and bridging the gap between experts in different fields. The book is intended for people of diverse academic and practical backgrounds and thought as a reference for researchers, engineers, and graduate students alike. Therefore, it will present a contemporary and comprehensive description of both theoretic abstraction and practical issues inherent in the distant ASR problem. The book has 14 chapters and covers: Overview of distant speech recognition, The sound field, Filter techniques, Speech feature extraction, Hidden Markov model parameter estimation, Search, Speech feature enhancement, Feature and model adaptation, Multi-microphone processing, Speaker localization and tracking, Blind source separation, Beamforming and Applications
- ISBN: 978-0-470-51704-8
- Editorial: John Wiley & Sons
- Encuadernacion: Cartoné
- Páginas: 594
- Fecha Publicación: 17/04/2009
- Nº Volúmenes: 1
- Idioma: Inglés