Title :
Sub-band weighted projection measure for sub-band speech recognition in noise
Author :
Nasersharif, B. ; Akbari, A.
Author_Institution :
Dept. of Comput. Eng., Iran Univ. of Sci. & Technol., Tehran, Iran
fDate :
7/6/2006 12:00:00 AM
Abstract :
A model adaptation method, named sub-band weighted projection measure, is proposed. This method is used to adapt concatenated sub-band feature vectors to hidden Markov model Gaussian mean vectors in noisy conditions. The method significantly improves the sub-band speech recognition rate in the presence of additive noise.
Keywords :
Gaussian noise; hidden Markov models; speech recognition; Gaussian mean vectors; additive noise; hidden Markov model; sub-band speech recognition; sub-band weighted projection measure;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20061222
