Robust Eye Features Extraction Based on Eye Angles for Efficient Gaze Classification System

Detection of eye gaze direction is a hot topic for research in the computer vision area which can be used in many applications. Although significant eye tracking techniques have been presented by the researchers for the last years, it is still the challenging task for improving the performance of the gaze detection system. This paper presents a new eye feature extraction system to build a robust eye gaze classier which uses the Viola-Jones algorithm to face detection and Constrained Local Neural Field model for eye region localization. Furthermore, geometry features of the eye are extracted from the detected eye region based on angles of a triangle of the eye. The algorithms were tested by a new dataset created from 34 participant females and males in different ages. The experimental results show that this method has better features extraction for the classification process.