Eye Fatigue Detection using Machine Learning and Deep Learning Classifiers

Amid the rising occurrence of visual fatigue associated with the widespread utilization of electronic displays, particularly among children using mobile phones and laptops, we have initiated the creation of a virtual model. Harnessing artificial intelligence and advanced learning methodologies, our model utilizes Kaggle datasets and employs effective algorithms to train four distinct models for detecting human eye fatigue. Determining the most precise model relies on individual training and testing accuracies. The optimal model is then integrated with a live webcam input to evaluate new, unseen data for predicting eye fatigue. Factors like pupil accommodation, blink frequency, eye-closed duration, and ocular size can be taken into account for the analysis; however, our primary focus is on ocular size, a crucial criterion.The chosen model undergoes hyperparameter tuning to enhance performance. Subsequently, the optimal model is connected to hardware components such as the Wi-Fi module ESP8266 and an electronic relay signal circuit for real-time applications. Following careful tuning and enhancements, the final model is ready for real-world application, interfacing with computer systems through components like the NodeMCU and an electronic relay signal circuit. This integration empowers the model to control essential functions, such as alerting users with cautionary messages and shutting down desktop monitors, promoting the protection of users' health and well-being.