Publication Details
QUALCOMM-ICSI-OGI Features for ASR
Dupont Stephane, Msc
Garudadri Harinath, Mcs.
Grézl František, Ing., Ph.D. (DCGM)
Heřmanský Hynek, prof. Ing., Dr. Eng. (DCGM)
Jain Pratibha, Msc.
Kajarekar Sachin, Msc.
Morgan Nelson, prof.
feature extraction, distributed speech recognition, Aurora task, RASTA, TRAP
Feature extraction module for the Aurora task based on a combination of a Wiener filtering with linear discriminant RASTA filtering and nonlinear tempoRAl Pattern (TRAP) classifier.
Our feature extraction module for the Aurora task is based on a combination of a conventional noise supression technique (Wiener filtering) with our temporal processing technigues (linear discriminant RASTA filtering and nonlinear TempoRAl Pattern (TRAP) classifier). We observe better than 58% relative error improvement on the prescribed Aurora Digit Task, a performance level that is somewhat better than the new ETSI Advanced Feature standard. Furthermore, to test generalization of our approach to an independent test set not available during development, we evaluate performance on American English SpeechDatCar digits and show 10.54% relative improvement over the new ETSI standard.
@inproceedings{BUT10471,
author="Lukáš {Burget} and Stephane {Dupont} and Harinath {Garudadri} and František {Grézl} and Hynek {Heřmanský} and Pratibha {Jain} and Sachin {Kajarekar} and Nelson {Morgan}",
title="QUALCOMM-ICSI-OGI Features for ASR",
booktitle="Proc. 7th International Conference on Spoken Language Processing",
year="2002",
pages="4--7",
publisher="International Speech Communication Association",
address="Denver",
isbn="1-876346-42-6",
url="http://www.fit.vutbr.cz/~burget/phd_activities/adami_icslp02.pdf"
}