On the Use of Grapheme Models for Searching in Large Spoken Archives
Jan Svec, Josef V. Psutka, Jan Trmal, Lubas Smfdl, Pavel Ircing, Jan Sedmidubský
Abstract
This paper explores the possibility to use grapheme-based word and sub-word models in the task of spoken term detection (STD). The usage of grapheme models eliminates the need for expert-prepared pronunciation lexicons (which are often far from complete) and/or trainable grapheme-to-phoneme (G2P) algorithms that are frequently rather inaccurate, especially for rare words (words coming from a different language). Moreover, the G2P conversion of the search terms that need to be performed on-line can substantially increase the response time of the STD system. Our results show that using various grapheme-based models, we can achieve STD performance (measured in terms of ATWV) comparable with phoneme-based models but without the additional burden of G2P conversion.
BibTeX
@inproceedings{icassp2018_ontheuseofgraphe,
title = {On the Use of Grapheme Models for Searching in Large Spoken Archives},
author = {Jan Svec and Josef V. Psutka and Jan Trmal and Lubas Smfdl and Pavel Ircing and Jan Sedmidubský},
booktitle = {ICASSP 2018},
year = {2018}
}