Self-Applicable Eye Strain Detection Through the Measurement of Blink Rate Using Raspberry Pi
Mon Timothy Isaac M. Carcellar, Carl Justin S. Tychuaco, Analyn Niere Yumang · 2024
You Only Look Once (YOLO) is a Convolutional Neural Network (CNN) used primarily for object detection models due to its high accuracy and reliable processing speed. In the study, YOLO was used to calculate the blink rate of an individual through a twenty (20) second recording. The model utilized YOLOv5, which was trained using one hundred (100) photos of faces obtained from the public dataset CelebFaces Attributes (CelebA). The system used a Raspberry Pi 3B+ and a 5MP Camera Module for video recording. Testing the model on 40 respondents (20 seconds each with 160 frames) yielded an overall accuracy of 85% when assessed using a confusion matrix. The paper benefits individuals who spend an increasing amount of time in front of a computer screen by helping them evaluate the stress level of their eyes. The results of the study can be used for future monitoring devices of eye strain.