Real Time Eyeball Tracking via Derivative Dynamic Time Warping for Human-Machine Interface

In this paper, a real time, user independent eyeball tracking approach is presented. The system is implemented by using a low cost webcam. The robustness of the system is measured by several criteria such as users of different age where some of the users are wearing glasses under varying lighting condition, pose, eye orientation and distance from camera. The size and location of the region of interest which contains both eyes are made adaptive. Derivative Dynamic Time Warping is chosen as the classifier for this experiment since it can match patterns from data sequences with different lengths. Finally, the results, advantages, limitations and future works of the proposed method are reported. The online eye tracking procedure shows good accuracy and robustness when processing online image sequences at 50 frames/s on a 253 GHz Pavilion DV4 HP notebook