Title :
Recognition of pathological voices
Author :
Salma, Chekili ; Asma, Belhaj ; Aicha, Bouzid ; Noureddine, Ellouze
Author_Institution :
Lab. de Signal, Images et Technol. de l´Inf., Univ. Tunis el Manar, Tunis, Tunisia
Abstract :
In this paper, we are interested in recognizing normal and pathological voices using the classification method SVM. In This classification we add to the feature vector composed of MFCC coefficients, delta, delta second, Energy, fundamental frequency, a new parameter called the open quotient. The open quotient is defined as the ratio of the open phase by the mean pitch period. This parameter is calculated using the multiscale product method (MPM). SVM classification is then applied on the Kay Pentax database.
Keywords :
feature extraction; speech recognition; support vector machines; Kay Pentax database; SVM; feature vector; mean pitch period; multiscale product method; open phase; open quotient; pathological voices recognition; Databases; Mel frequency cepstral coefficient; Pathology; MFCC; Open quotient; SVM; pathological voices;
Conference_Titel :
Multi-Conference on Systems, Signals & Devices (SSD), 2014 11th International
Conference_Location :
Barcelona
DOI :
10.1109/SSD.2014.6808900