Improvement of speech recognition by nonlinear noise reduction
Data Analysis, Statistics and Probability
2007-06-20 v5 Computational Physics
Abstract
The success of nonlinear noise reduction applied to a single channel recording of human voice is measured in terms of the recognition rate of a commercial speech recognition program in comparison to the optimal linear filter. The overall performance of the nonlinear method is shown to be superior. We hence demonstrate that an algorithm which has its roots in the theory of nonlinear deterministic dynamics possesses a large potential in a realistic application.
Cite
@article{arxiv.physics/0507044,
title = {Improvement of speech recognition by nonlinear noise reduction},
author = {Krzysztof Urbanowicz and Holger Kantz},
journal= {arXiv preprint arXiv:physics/0507044},
year = {2007}
}
Comments
see urbanowicz.org.pl