<![CDATA[Drowsiness during work can reduce productivity and increase the risk of errors, particularly in monotonous office jobs that rely heavily on screen-based activities. This study aims to design and develop a modular computer vision-based drowsiness detection system capable of detecting drowsiness conditions in real-time and providing automatic notifications to users. The system detects drowsiness conditions using the Eye Aspect Ratio (EAR) method through an internal laptop or PC camera. The implementation was carried out using the Python programming language with the OpenCV library for image processing and Dlib for facial landmark detection. Users are provided with a graphical user interface (GUI) application developed using Tkinter. The system automatically sends warning messages through the Telegram Bot API when the EAR value is detected below the threshold of 0.21 for more than 3 seconds. Notifications can be delivered automatically to both the user’s and supervisor’s Telegram accounts. The testing results indicate that the system is capable of operating in real-time with fast response times and stable notification delivery through Telegram. Warning messages are delivered concisely so as not to disrupt the user’s workflow. Based on these findings, the developed drowsiness detection system has the potential to be utilized as a supporting tool to improve alertness and prevent productivity decline among office workers.]]>
Copyrights © 2026
