Remote eye gaze tracking system as a computer interface for persons with severe motor disability

Abstract

State of the Art human computer interfaces (HCI) for assisting individuals with severe motor disabilities employ remote eye-gaze tracking (EGT) systems which obtain eye coordinates and convert them into mouse-pointer coordinates. The performance of those systems is traditionally affected by mouse-pointer jitter and miscalibration due to head movement. This study addresses this problem and proposes an interface to minimize those errors. The interface allows inspecting and quantifying those errors and collecting the necessary information which is used in real time for training an artificial neural network which improves the coordinate conversion mechanism. The novelty of this study resides in the integration of several procedures, such as: (a) error inspection at system startup, (b) real time improvement of the eye-to-mouse-pointer coordinate conversion mechanism, (c) determination of a practical solution to the mouse click operations, and (d) development of effective means to monitor and evaluate the system performance. © Springer-Verlag 2004.