Drowsiness Detection Using Heart Rate Variability and Facial Expression

Teh, Wei Chong (2025) Drowsiness Detection Using Heart Rate Variability and Facial Expression. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text Teh Wei Chong -FULL TEXT.pdf Restricted to Registered users only Download (10MB)

Abstract

Drowsy driving is a major road safety issue, contributing to numerous accidents and fatalities worldwide. This paper aims to develop a real-time drowsiness detection system that integrates facial expression analysis to enhance driving safety. The proposed system will analyse eye closed open and yawning as parameters to detect patterns of driver tiredness and trigger alerts to prevent potential accidents and the parameter is based on the output class label predicted by YOLO model training. The methodology includes gathering and preprocessing data from multiple sources, then performing feature extraction, integration, and validation to ensure reliable performance across different conditions. YOLOv8 and YOLOv11 are utilized to evaluate and determine the most suitable model. Detecting sleepiness in a timely and accurate manner is the core focus of this research, with an emphasis on balancing detection precision and real-time responsiveness. The ultimate goal is to build a system that delivers prompt alerts to improve road safety. Future work will involve implementing and refining these models to ensure efficient drowsiness detection and contribute to overall traffic safety.