Eye Tracking Analysis Using Convolutional Neural Network

Abstract

Eye tracking plays a pivotal role in fixing the user interface allowing one to understand what a person is actually looking at while browsing through a webcam or external cam or using an infrared eye tracker etc., which depends on needs and conditions. With eye-tracking, one can easily test any video material, AD performance, package design concepts, product shelf placement, website performance, mobile websites, and apps. It can be the supreme technology in providing various insights into the processes which involve application into various fields of academics, science & technology, marketing, and other researchers. The goal of eye tracking is to detect and measure the point of gaze (where one is looking) or the motion of eye(s) relative to the head. This study examines the current state-of-the-art in deep learning-based gaze estimation algorithms, with a particular focus on Convolutional Neural Networks (CNN). Several studies are focusing on various approaches for dealing with different head pose and gaze estimation. Large-scale gaze estimate datasets with various head poses and illumination conditions were reported in the current study. We are building a model detecting if the eyes captured are right or left and detecting the gazing point and the aim is to solve the problem if they are accurate. This defined problem requires a method with high learning capacity which is able to manage the complexity of the given dataset. For the present study Convolutional neural network(CNN) has proved effective to get better results for the defined problem.