Ebook: Automatic speech recognition on mobile devices and over communication networks

The remarkable advances in computing and networking have sparked an enormous interest in deploying automatic speech recognition on mobile devices and over communication networks. This trend is accelerating.

This book brings together leading academic researchers and industrial practitioners to address the issues in this emerging realm and presents the reader with a comprehensive introduction to the subject of speech recognition in devices and networks. It covers network, distributed and embedded speech recognition systems, which are expected to co-exist in the future. It offers a wide-ranging, unified approach to the topic and its latest development, also covering the most up-to-date standards and several off-the-shelf systems.

Key features:

• Provides an in-depth review of network speech recognition, distributed speech recognition, embedded speech recognition, systems and applications

• Begins with a comprehensive overview of the subject, discussing the pros and cons of the presented approaches, and guiding the reader through the following chapters

• Includes platforms like mobile phones, PDAs and automobiles

• Presents state-of-the-art methods, advanced systems, and the latest standards

• Offers working knowledge needed for both research and practice

• References supplemental material at associated complementary website at: http://asr.es.aau.dk

This all-inclusive text/reference is an essential read for graduate students, scientists and engineers working or researching in the field of speech recognition and processing. It offers a self-contained approach to this hot research topic.

---

As cellphones become ever more popular, and computational costs continue to fall, the prospect of ASR over wireless networks is attractive. So too is the use of ASR over the Internet (VoIP). The book describes the current state of understanding of both these ideas. A common configuration is for the speech to be captured and digitised on a client machine [eg. cellphone], and the signal then sent over a network to a server on which runs the ASR. Several speech recognition standards have arisen over the years, to quantify the digitising and the ASR effectiveness. For the ASR, Hidden Markov Models appear to be commonly used. The efficacy is described. Maybe the most advanced topic is speech to speech translation, via handheld devices. Seriously difficult. Not only are there the problems of ASR accuracy for the input speech, but then the well known problems of Machine Translation of that to text in another language.