Methods of Control Improvement in an Eye Tracking Based Human-Computer Interface

In this paper, different optimization techniques for pupil detection algorithms used in eye tracking applications are comparatively analyzed. Due to different noise sources which affect the eye images captured with an infrared video camera, a large noise component is overlapped on the signals provided by the pupil detection algorithms. This noise component determines cursor instability on the user screen, which makes difficult the eye tracker operation. In order to increase the cursor stability on the user screen, different techniques based on real-time filtering, high frequency spikes cancelling from the signals provided by the PDA and the snap-to-point technique can be used. The experimental results performed with the improved version of the PDA confirm the significant increase of the eye tracker system performances.