Effective data conversion algorithm for real-time vision based human computer interface

  • Authors:

  • Anaelis Sesin;Malek Adjouadi;Melvin Ayala

  • Affiliations:

  • Department of Electrical & Computer Engineering, Florida International University, Miami, FL;Department of Electrical & Computer Engineering, Florida International University, Miami, FL;Department of Electrical & Computer Engineering, Florida International University, Miami, FL

  • Venue:
  • SEPADS'07 Proceedings of the 6th WSEAS International Conference on Software Engineering, Parallel and Distributed Systems
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Human computer interfaces (HCI) for assisting persons with disabilities may employ eye gazing as the primary computer input mechanism. These systems rely on the use of remote eye-gaze tracking (EGT) devices to compute the direction of gaze and employ it to control the mouse cursor. Regrettably, the performance of these interfaces is traditionally affected by inaccuracies inherited from the eye tracking devices and ineffective EGT to mouse-pointer data conversion mechanisms. This study addresses this problem and proposes a new optimized data conversion mechanism. It analyzes in more details the correlation between the two data types resulting in a considerable increment in the accuracy of the system. This improved data conversion interface integrates the following procedures: (a) map the correlation between the EGT data and the mouse cursor position, (b) apply a curve fitting method that best suits the behavior of the data, (c) interpret the direction of gaze in order to determine the appropriate mouse cursor response, and (d) use effective means to monitor and evaluate the system performance.