Drowsiness Detection in Virtual Meetings Using Machine Learning

Abstract

"This thesis explores the development and implementation of a drowsiness detection system integrated within virtual meeting platforms, utilizing machine learning and computer vision techniques to enhance meeting productivity by monitoring participant alertness. The performance of traditional drowsiness detection methods, which typically require substantial manual intervention and are not tailored for virtual environments, is significantly limited by these factors. In response to these challenges, this study introduces an innovative drowsiness detection model that automates the detection process using Convolutional Neural Networks (CNNs) and real-time video analysis. This system is designed to identify signs of drowsiness through facial cues such as eye closure and yawning, extracted from video feeds of meeting participants. The effectiveness of the proposed system was rigorously tested in diverse lighting conditions and across various user demographics to ensure accuracy and reliability. The system's performance was benchmarked against traditional models to demonstrate its superior efficiency and effectiveness. Initial tests on simulated meeting scenarios show that the drowsiness detection system not only improves detection accuracy but also operates seamlessly within the virtual meeting framework, thereby maintaining user engagement without significant disruptions to meeting flow. The outcomes of this research indicate that integrating drowsiness detection into virtual meetings can significantly mitigate the impacts of reduced participant alertness, thus enhancing the overall productivity and effectiveness of remote collaborations. This thesis lays the groundwork for future advancements in remote monitoring technologies and opens new avenues for research into behavioral analytics within virtual environments. "