Facial feature tracking for eye-head controlled human computer interface

In this paper, we propose a robust fast and cheap scheme for locating the eyes, lip-corners, and nostrils for eye-head controlled human computer interface on a facial image with non-constrained background. Many researchers have presented eye tracking methods. But the methods are not both robust and fast, and they also have many limitations. The method we propose uses complete graph matching from thresholded images. That is, after labeling the binarized image that is separated by a proper threshold value, the algorithm computes the similarity of between all pairs of objects. After that, the two objects that have the greatest similarity are selected as eyes. The average computing time of the image (360*240) is within 0.2 (sec) and if the search window is reduced by estimation, the average computing time is within 0.1 (sec). It has been tested on several sequential facial images with different illuminating conditions and varied head poses. It returned quite a satisfactory performance in both speed and accuracy. The algorithm is highly cost effective.