Speech Summarization
Yang Liu, Dilek Zeynep Hakkani-Tür · 2011
The increasing amount of audio recordings has attracted a lot of research interest in speech summarization. This chapter covers some techniques that have been developed for speech summarization. It describes task definition, evaluation issues, and some available resources, as well as discusses some issues and future directions. A speech summarization system typically involves addressing the following problems: speech recognition, summarization, and rendering the output. Speech summarization can be a useful tool to handle the overwhelming amount of information available in audio recordings. One straightforward approach for speech summarization would be to applymature text summarization techniques to the speech transcriptions. Indeed, some text summarization techniques have been successfully used or adapted to deal with speech input. Dramatic reduction in word error rate (WER) has been achieved in speech recognition during the past decades. The chapter describes some of the data resources that have been developed. Controlled Vocabulary Terms audio recording; error statistics; speech recognition