Eye Movement Tracking for Computer Vision Syndrome using Deep Learning Techniques

Due to the increased usage of digital devices in daily life, particularly among children, symptoms such as drying of the eyes, eye strain, headaches, blurred vision, etc., have become recurrent nowadays. Extensive use of computers and smartphones may lead to a common eye-related condition known as Computer Vision Syndrome (CVS). It is often characterized by a reduced blinking rate of the user. In this paper, we propose a deep neural network and computer vision-based machine learning model that entails training a Convolutional Neural Network (CNN) to detect eye blinks, and monitoring blink rates with a Long Short-Term Memory (LSTM) network. This model can be incorporated into smartphones and computers in the form of background apps and may help prevent the risk of CVS or similar disorders. Inferences about the blink rate and eye movement patterns have also been identified. Our model is implemented using TensorFlow and Dlib libraries and has been trained on the Closed Eyes in the Wild (CEW) dataset. The network achieved an accuracy of 94.2% when trained on non-RGB images of eye patches and 91.4% on RGB facial images in real-time.