Developing A Video-Based Discussion Platform With Emotion Detection Deep Learning To Increase Online Learning Engagement

Abstract

This study discusses the development of Videmo Platform, an online learning platform that integrates Video-Based Discussion (VBD) and Emotion Detection Deep Learning (EDDL) technology to promote deeper learning. Videmo Platform is a response to the limitations of text-based discussion, which often miss out on capturing the students' emotions needed for interaction and understanding. Videmo recognises emotional expressions such as happy, sad, angry and surprise using EDDL technology through facial analysis in videos discussion. Technical test results showed that the platform achieved an F1-score detection accuracy of 95% for MobileNet V1 SSD, followed by Tiny Face Detector F1-score of 92.0% and Face Expression Model F1-score of 88.0% in various lighting conditions and viewing angles. Surveys with 34 students found 86.27% of students said that Videmo engaged them more in the discussion and 75% of lecturers felt they understood the students’ emotional engagement better. These findings illustrate the ability of Videmo Platform to transform the online learning experience into one that is more interactive and personalized.