An Eye Gaze Tracking System Using Customized User Profiles to Help Persons with Motor Challenges Access Computers

The central aim of this study is to develop an adaptive real-time eye-gaze tracking (EGT) system that serves as an assistive tool for persons with motor challenges access computers with optimal practicality. The novelty of the proposed method is that it adapts to the different and changing jitter characteristics of each specific user, through the configuration and training of an artificial neural network (ANN). A profile, generated for each user by the ANN through a one-time short training session, comes to reinforce the stability of mouse-cursor movements. The user profile, which can be fine-tuned with additional training, and the methods for training are embedded in the proposed system. The results using 9 subjects show an average jitter reduction of 36% in test 1 which is to follow a moving target, and 53% for following the contour of a square, which resulted in eye-gaze displacements that are significantly smoother.