FEATURES EXTRACTION AND TRAINING STRATEGIES IN CONTINUOUS SPEECH RECOGNITION FOR ROMANIAN LANGUAGE

Corneliu Octavian Dumitru, Inge Gavat

2006

Abstract

This paper describes continuous speech recognition experiments for Romanian language, by using HMM (Hidden Markov Models) modeling. The following questions are to be discussed: the realization of a new front-end reconsidering linear prediction, the enhancement of recognition rates by context dependent modeling, the evaluation of training strategies ensuring speaker independence of the recognition process without speaker adaptation procedures, by speaker selection for training. The experiments lead to a development of the initial system with a promising front-end based on PLP (Perceptual Linear Prediction) coefficients, second ranked for the recognition performance obtained, near the first ranked front-end based on mel-frequency cepstral coefficients (MFCC), but far better as the last ranked, based on simple linear prediction. Concerning the implemented algorithm for context dependent modeling, it permits in all situations enhanced recognition rates. The experiments made with gender speaker selection enhanced under certain conditions the recognition rate, proving good generalization properties especially by training with the male speakers database.

Paper Citation

in Harvard Style

Octavian Dumitru C. and Gavat I. (2006). FEATURES EXTRACTION AND TRAINING STRATEGIES IN CONTINUOUS SPEECH RECOGNITION FOR ROMANIAN LANGUAGE . In Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics - Volume 3: ICINCO, ISBN 978-972-8865-61-0, pages 114-121. DOI: 10.5220/0001198901140121

in Bibtex Style

@conference{icinco06,
author={Corneliu Octavian Dumitru and Inge Gavat},
title={FEATURES EXTRACTION AND TRAINING STRATEGIES IN CONTINUOUS SPEECH RECOGNITION FOR ROMANIAN LANGUAGE},
booktitle={Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics - Volume 3: ICINCO,},
year={2006},
pages={114-121},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001198901140121},
isbn={978-972-8865-61-0},
}

in EndNote Style

TY - CONF
JO - Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics - Volume 3: ICINCO,
TI - FEATURES EXTRACTION AND TRAINING STRATEGIES IN CONTINUOUS SPEECH RECOGNITION FOR ROMANIAN LANGUAGE
SN - 978-972-8865-61-0
AU - Octavian Dumitru C.
AU - Gavat I.
PY - 2006
SP - 114
EP - 121
DO - 10.5220/0001198901140121