Drowsiness detection and alert system using face recognition with Raspberry Pi / Nur Syazwina Mohd Fadzalisham, Nor Adora Endut and Muhammad Nizamuddin Rosli

• Main Authors: Mohd Fadzalisham, Nur Syazwina, Endut, Nor Adora, Rosli, Muhammad Nizamuddin
• Format: Book Section
• Language: English
• Published: Faculty of Computer and Mathematical Sciences 2023
• Subjects: Integer programming
• Online Access: https://ir.uitm.edu.my/id/eprint/94301/1/94301.pdf; https://ir.uitm.edu.my/id/eprint/94301/

Recently, drowsiness detection has garnered significant attention due to its crucial implications in various industries, such as transportation, healthcare, and workplace safety. Drowsiness, often caused by exhaustion or lack of sleep, poses serious risks to people's safety and well-being. Even in less critical situations, like during online classes, feeling sleepy negatively impacts students' learning and academic performance. The occurrence of accidents and errors resulting from drowsiness has underscored the need for effective detection technologies to mitigate these risks. The main objective of this project is to develop a drowsiness detection and alert system using a Raspberry Pi. The technology aims to analyse facial features in real-time, efficiently identifying key markers of drowsiness through computer vision techniques and machine learning algorithms. Leveraging Raspberry Pi as the camera component offers a portable and cost-effective solution suitable for various settings. The solution integrates a Telegram bot for streamlined communication, utilizing Pi Camera to capture facial photos and promptly detect signs of drowsiness. This bot sends rapid alert messages to users' mobile phones or laptops, enabling swift responses to any concerns related to potential drowsiness, thereby enhancing safety and well-being. The system also proficiently records essential drowsiness data in a MySQL database, allowing for further analysis and insights to continuously improve and enhance effectiveness in reducing drowsiness-related incidents.