During this project we have carried out interviews with speech and occupational therapists in order to better understand the limitations of current vision-based eye gaze tracking systems and to obtain feedback on the proposed EOG-based system. Through these meetings, we have identified the most popular virtual keyboard layouts and incorporated them into our EOG-based system. Furthermore, we have also developed a symbol-based communication menu intended to help younger individuals with language disabilities to communicate.
Through the analysis of the manifestation of eye movements in the electrooculogram, this project has developed algorithms which can detect eye movements in real time and translate these movements to pixel coordinates on a computer screen. The latter enables a user to select icons on a computer screen using eye gaze and hence operate an application without the need of keyboards or mice. In order to select an icon, users fixate on the desired key for a predefined period of time, known as a dwell-time, to activate the corresponding function. Normally this dwell-time ranges from 0.8 to 2 seconds.
In order to provide a natural mode of interaction, the team of researchers have also developed methods which compensate for head movements such that users do not need to maintain a fixed head pose while using the developed EOG based applications. The system takes into consideration the dynamic aspects of the eye-head coordination that normally take place when one shifts their point of gaze (POG) form one target to another and incorporates this information into the system to ensure that accurate POG estimates can be obtained under such natural movements.
