The Identification and Classification of Facial Expressions under Stress using an EMG

Due to the growing number of people suffering from stroke or facial nerve damage which influence the quality of life, we propose an acquisition system that identify and classify facial expression by manual analysis of signals using an EMG. Beside these features, the designed system is able to save the records in a database so that these kinds of medical conditions can be seen when they first appeared and how they manifested along the way. The acquisition system has an interface created in MATLAB App Design, which facilitates the experience of signal acquisition and database creation. The signal acquisition occurred during the visual or auditory stimulation of four reactions, by means of videos made for this purpose. The extracted signals were manually segmented and analyzed. For that acquisition system, a cover was designed, and 3D printed, which has both a protective role and an aesthetic role. Although the paper investigates only a few facial expressions, it can be extended in the future to recognize emotions and even to make correlations between different facial expression and other medical conditions recorded in the database.