An Efficient Algorithm To Collect Minimal Speech Corpora

Saad Irtza; Sarmad Hussain

No Detail Found!

An Efficient Algorithm To Collect Minimal Speech Corpora

Author(s)

Saad Irtza Electrical Engineering Department, University of Engineering & Technology,Lahore,Pakistan

Sarmad Hussain Center for Language Engineering, KICS, University of Engineering & Technology, Lahore, Pakistan

Abstract

Generally phonetically rich and balanced corpora are popular for training speech recognition system but these corpora are costly to develop. Different greedy algorithms have been develop to collect such corpora. A significant effort is required to record and transcribe such speech corpora. Therefore there is motivation to further reduce their size. This paper demonstrates such an algorithm. Earlier work shows that different amount of training data is required to train different phonemes. The current work further develops these findings to reduce phonetically rich training data. Experiments show that this algorithm reduces the size of an Urdu speech corpus by 56.49% without degradation in accuracy.

Publication Details

Page(s) 11-18

DOI DOI not available

Published Journal: Pakistan Journal of Engineering and Applied Sciences, Volume: 17, Issue: 0, Year: 2015

Request the full text from PASTIC

Keywords

Urdu speech corpus, Reduced speech corpus

Citations

Downloads

Views