Eye Detection and Face Recognition using Line Edge Maps

The task of face recognition has been actively researched in recent years , and many techniques have been proposed. This paper reports on the design, implementation and experimental study of an improved technique for face recognition, including eye detection in the preprocessing phase of a localized face image. After a brief summary of some known face recognition techniques, we present an innovative algorithm for eye detection on gray intensity face image, which combines a feature-based approach using line edge map (LEM), geometrical properties of eyes, and the use of LEM eye templates. Matching in each case is carried out using primary line segment Hausdorff distance. Experimental results of face recognition using this technique are presented. A potential enhancement is also proposed, namely facial feature block-wise implementation of the LEM technique. Index Terms – Line edge map, Eigenface, Bayesian face recognition, Eye detection, Line segment Hausdorff distance.