Speech quality measurement automation for patients with cancer of the oral cavity and oropharynx

This work is dedicated to the measurement of the quality of the speech signal in the process of rehabilitation after surgery in the treatment of cancer of the oral cavity and oropharynx. The current state of the problem and approaches to its solution used at the Tomsk Cancer Research Institute are analysed. Disadvantages of these approaches and the task aimed at addressing these shortcomings are presented. The phonemes groups at greatest change in the surgical treatment of cancer of the tongue are determined. A comparison of the data with the classical works and shows the applicability of the approach used and the correctness of its implementation are presented. A preliminary analysis of the spectrograms before and after surgery is showed. Is made the assumption that exposure to changes in the first place the upper part of the spectrum. Is identified the tasks for the next phase of work.