The paper introduces a new class of signal processing techniques that are trained on large amounts of speech data to extract a set of features for automatic recognition of speech. Such optimized signal processing techniques appear to be consistent with some important properties of human hearing.
Cite as: Hermansky, H. (2003) Data-guided processing of speech. Proc. Workshop on Spoken Language Processing, 15-21
@inproceedings{hermansky03_wslp,
author={Hynek Hermansky},
title={{Data-guided processing of speech}},
year=2003,
booktitle={Proc. Workshop on Spoken Language Processing},
pages={15--21}
}