A Supervised Fuzzy Eye Pair Detection Algorithm
Seba Susan, Pooja Kadyan · 2013
Most of the face recognition problems in literature rely on automatic eye-pair detection algorithms for locating the eye positions followed by the normalization of the face image based on the distance between the eyes. In this paper we propose a supervised fuzzy eye pair detection algorithm that can be executed in real time and requires minimal training. Nine categories of facial geometrical measurements are defined. The Gaussian function is used to compute the fuzzy memberships with the mean and standard deviation of the Gaussian being evaluated from ten reference images from the database chosen randomly. The eye pair detection algorithm works successfully on the Utrecht face database except for two cases where the eyebrow pairs are detected. The results are shown to outperform the popular eye variance filter method for eye detection.