Automated 20-20-20 Timers Using Facial Detection and Facial Recognition
Saumodip Das, Geetanjali Pal, Somsubhra Manna, Arnab Dhara, Anindya Sen · 2023
This paper proposes two automated 20-20-20 timers using facial detection and facial recognition respectively. Computer Vision Syndrome (CVS) is a common phenomenon these days, resulting in eye irritation and headaches. A possible prevention of CVS is adopting the 20-20-20 rule. Two types of 20-20-20 timer exist in the market, but both have scopes of improvement. While one type of timer needs to be manually set by the user like a stopwatch, the other type of timer detects the ON -screen time of monitors as a parameter. Both these models fail to detect whether the user is actually looking at the screen, and interrupt the work schedule of the users. The authors have developed a smart automated 20-20-20 timer that uses haar-cascade classifiers for detecting the face and accordingly adapt to the situation. This timer is able to detect the facial features in real time, and alert the user after 20 minutes of screen time. Furthermore, the authors have developed another smart 20-20-20 timer using YOLOv5s. This timer can recognise multiple users looking at the screen and run the timer independently for each person. While the haar-cascade enabled timer is preferred for homes and personal laptops, desktops and PCs; the YOLOv5s enabled timer is suited for offices, schools, colleges and libraries. This paper proposes an automated smart 20-20-20 timer system.