An adaptive template eye location based on Gabor transform method using web camera

An adaptive template matching method for human eye location is proposed by using ordinary web camera. Face video is captured by using an ordinary web camera. The face region is firstly detected by applying face detection algorithm based on Haar-like feature. The pupil position in human face frame is determined based on Gabor transform eye location method, which is used as an initial eye template. The following eyes are located by matching the pupil in previous frame as adaptive template. Experimental results show that the proposed eye location method is more accuracy than the traditional eye template matching method.