Improving mobile device interaction by eye tracking analysis

This paper describes a non-intrusive eyetracking tool for mobile devices by using images acquired by the front camera of the iPhone and iPod Touch. By tracking and interpreting the user´s gaze to the smartphone´s screen coordinates the user can interact with the device by using a more natural and spotaneous way. The application uses a Haar classifier based detection module for identifying the eyes in the acquired images and subsequently the CAMSHIFT algorithm to find and track the eyes movement and detect the user´s gaze. The performance of the proposed tool was evaluated by testing the system on 16 users and the results shown that in about 79% of the times it was able to detect correctly the users´ gaze.