Prosodic Featured Based Automatic Language Identification  (Paperback, Niraj Singh)

Living beings inherently have the ability to differentiate languages as a part of human intelligence. Language Identification (LID) had been a science fiction in 1970’s but today; it has been deployed in practical usage. The prosodic features of a speech are relatively simpler in their structure and are accredited to be very affective in some Language Recognition (LR) or LID tasks; irrespective of these features to be biased on numerous factors, as speaker’s way of speaking, the culture and background of speaker. The book includes a series of experiments on several speech corpus with different classification or/and identification technique. At the end of each chapter, few review questions have been included and at the verge of the book, a short list of projects for research scholars has been mentioned in addition to a set of MCQs and Important questions. This book motivates for developing a multilingual LID system which can be widely used for betterment of mankind, particularly in the fields of Intelligence Police/Military) services and medical care. In an overview, we may assert that the book explores various experimental datasets, for, performance analysis of LID system with News speech and Natural Conversation speech; Joint Factor Analysis for LR on prosodic featured models and for automatic LID using i-Vector based prosodic system.