Classifying Emotions of Digital Immigrants while using Software Application based on Facial Features

In today’s age, the use and access to technology have become essential aspects of every individual. However, digital immigrants, individuals raised before the digital age, need help to cope with this advancement. In this study, the Education Technology Office conducted an onsite learning intervention to teach the digital immigrant faculty members about using Microsoft Applications such as Office365, Microsoft Teams, and Microsoft Forms. The faculty members’ facial expressions during the intervention were recorded using a web camera. Five consented videos of 1 hour were obtained in the data collection process. The video dataset was processed using the OpenFace toolkit to detect the facial features per frame. Afterward, feature selection and extraction of Action units were applied before fitting to the LSTM model. The LSTM model yielded a recognition accuracy of 97.3% on the validation set. In the analysis of facial expressions based on time, it was observed that as time progressed during the learning intervention, there were not many changes in the emotion of the participants. Also, the most frequent emotion recognized in different timestamps (initial, middle, final) part was ’happy,’ which may indicate that the participant has enjoyed the intervention and may suggest that it is effective. While other emotions were also evidently recognized, such as ’sad’ and ’surprised’, which might indicate the difficulty or challenges encountered by the digital immigrants during the learning intervention.