Study and realization of a tele-medical platform dedicated to the objective assessment and to the follow-up of dysphonia chronic laryngeal origin by spectro-temporal analysis of acoustic vocal signal

Abstract

In this paper we develop a system dedicated to the objective characterization of dysphonia chronic laryngeal origin. The purpose of this system is threefold: diagnosis, treatment and monitoring of patient. He includes a chain tele medical made up of: 1- Hardware with a microphone and a sound card of terminal computer. 2- a software with vocal Audacity environment for the recording of the acoustic signal, Wamp environment for the establishment of data base on line and Visual BASIC environment for the transfer of the data under protocol TCP/IP making profitable architecture customer server supported by the component Winsock of VB in relation to the operating system Windows. For that we design an experimental protocol which consists of the recording and the archiving of the acoustic vocal signal by means of the software environment Audacity which makes it possible to deliver a temporal signal under format WAVE. Our contribution consisted of the implementation, under the Visual Basic environment, of an algorithm that allows performing the analysis of a spectro-temporal acoustic voiced speech signal in this case the vowel “a” sustained for three seconds. We applied this algorithm on six healthy subjects and six pathological subjects of whose four with cancer of the larynx, one with chronic laryngitis and one presenting an inflammatory polyp of the vocal cords. The results obtained show a variability of spectro-temporal characteristics between healthy and pathological subjects and prone pathological between them according to the nature of the lesion in particular with regard to spectral content evaluated by DFT-RD and the fundamental frequency F0 averaged over several frames of the voiced signal.