Entity

Time filter

Source Type


Vieira R.T.,Electrical Engineering Coordination Federal Institute of Education | Brunet N.,Electrical Engineering Coordination Federal Institute of Education | Costa S.C.,Electrical Engineering Coordination Federal Institute of Education | Correia S.,Electrical Engineering Coordination Federal Institute of Education | And 2 more authors.
Journal of Medical and Biological Engineering | Year: 2012

Laryngeal diseases usually affect vocal quality. An appropriate acoustic analysis of the voice can be used as an auxiliary non-invasive tool for the pre-diagnosis of laryngeal pathologies. It is possible to evaluate the effectiveness of some medical treatments, as well as pre or post-surgical patient evaluation. This work investigates the use of some speech signal features to discriminate pathological voices from healthy voices. To detect vocal disorders caused by Reinke's edema or vocal fold paralysis, an acoustic analysis is conducted using three entropy measures (Shannon, relative, and Tsallis) and four cepstral coefficients (cepstral, delta cepstral, weighted cepstral, and weighted delta cepstral). The performance of individual classifiers based on each measure is evaluated. Then, the measures are combined considering three rules: average, product, and weighted sum. Classification accuracy is improved when combinations of acoustic features are considered compared to using individual classifiers. Source


Tavares R.,Electrical Engineering Coordination Federal Institute of Education | Brunet N.,Electrical Engineering Coordination Federal Institute of Education | Costa S.C.,Electrical Engineering Coordination Federal Institute of Education | Correia S.,Electrical Engineering Coordination Federal Institute of Education | And 2 more authors.
2011 ISSNIP Biosignals and Biorobotics Conference: Biosignals and Robotics for Better and Safer Living, BRC 2011 | Year: 2011

This work aims at investigating the combination of classifiers based on three different entropy measurements (Shannon, Relative and Tsallis) and four cepstral coefficients (cepstral, delta-cepstral, weighted cepstral and weighted delta cepstral) to discriminate pathological voices under Reinke's Edema from normal ones. The performance of the combined classifiers is evaluated considering the average and product rules. Classification accuracy is improved when combinations of acoustic features were considered, compared to individual classifiers results. © 2011 IEEE. Source

Discover hidden collaborations