Project DetailsThe project deals with the diagnosis of various voice pathologies related to the throat and vocal cords which today can only be diagnosed by a long and multi-stage process that includes listening to the patient's voice by an otolaryngologist specialist and then an invasive examination using special equipment. We assume that there is plenty of information about those pathologies in the voice recordings of the subjects, and therefore we wish to use them to design a simpler diagnosis procedure that is based on machine-learning algorithms.
In the first phase of the project, we carried out a binary classification of the subjects into sick/healthy, according to the voice recordings, and we reached a performance of 82% on the test set. Later we also tested the multi-class case, in which we try to diagnose the specific disease. We reached impressive results in the categories of "dystonia" and "scars on the vocal cords after surgery", with the latter category being challenging to diagnose even for otolaryngologists.
