“Speaker Trainer” Application: Speech Processing Subsystem

Public speaking is an integral part of diverse professions as well as of students’ routines. Nowadays, the assimilation of information by the audience heavily depends not only on the direct content of the report but also on the oratorical qualities of the speaker. The quality of public performance consists of various factors, so it is generally a backbreaking task to evaluate the quality manually. In addition, human perception is often subjective. Thus, creating an automated solution to improve image quality will greatly simplify the process of improving speech quality. This solution will allow users to independently track their progress and improve their skills. This paper proposes a comprehensive solution for evaluating the quality of a speech by creating an automated speech processing subsystem and server side of the mobile application, which will solve the aforementioned problem. The speech processing subsystem will allow users to define problem areas in the speaker's speech based on the available video recording of it and provide recommendations for improving the quality of the speech. The server side of the Android application will also be created. It will be responsible for authenticating the user, uploading and processing of video recordings, providing processed data for the client side, and interaction with the database (viewing, storing, editing, and adding data).