Publisher : Advances in Intelligent Systems and Computing
Campus : Bengaluru
School : School of Engineering
Department : Electronics and Communication, Mathematics
Year : 2018
Abstract : pThis paper aims to discuss the development of an isolated word recognizer for the Indian language Kannada. The word recognizer is built using HTK, which is based on HMM. The system is trained using triphone HMMs for Kannada words in open space environment from 10 speakers and tested using data from 4 speakers. This paper also gives a comparison of results between MFCC and LPCC techniques for four different test sets. © Springer Nature Singapore Pte Ltd. 2018./p