Minimizing word error rate in textual summaries of spoken language
NAACL 2000 Proceedings of the 1st North American chapter of the Association for Computational Linguistics conference
| Hi-index | 0.00 |
There are open source tools available for voice based desktop control. Unfortunately these tools often work with languages that have reliable open source large vocabulary acoustic models. This paper describes the steps needed to build a specific small vocabulary acoustic model with a reduced set of desktop control commands. It includes a review of the state of the art of open source voice technologies and open source desktop control tools. It also describes the experimental work in Spanish. This language is especially well suited for this task because 1) it does have a large vocabulary acoustic model that can be used as baseline and 2) there are specific regional peculiarities that give a insight on the benefits of building small vocabulary specialized acoustic models.