Zdrojový dokument:Intelligent systems applications in software engineering : proceedings of 3rd computational methods in systems and software 2019, Vol. 1
Název akce3rd Computational Methods in Systems and Software, CoMeSySo 2019 (10.09.2019 - 12.09.2019, Zlín)
Abstrakt:
The main purpose of this paper is to show identification possibilities of voice differences for people whose voice has been influenced by any kind of voice disorder. Introduction of common diseases of vocal cords or larynx is followed by a chapter including ordinary treatment techniques. Even if the surgery ends up well the voice production can be affected in some way. Doctors are mostly able to measure only limited amount of voice characterizing parameters. More precise analysis of subjects within predefined time intervals should lead to more specific results and may prove more efficient. This article presents a different scientific approach which is based on the voice parameterization and analysis. The only thing needed for this kind of research is obtaining of recordings of analyzed subjects (before surgery, soon after that and then for example 2 months later). These recordings can be processed using common audio processing methods and required variables are extracted and saved in form of so-called feature vectors. Some features are expected to change as the result of treatment. Some of used methods are similar to ordinary techniques or they have something in common, but it allows to measure and identify even more variables describing the voice. Diagnostic experience can be supplemented by our software, where many parameters are visualized. But the final decision is still up to the doctor.