Unconstrained Face Recognition using Bayesian Classification

In this paper we propose a method for person identification. The proposed method is invariant to illumination, scale, pose, camera exposure and translation of the head. In order to make the model illumination invariant, a linear transform is applied. Binary affine features are used to extract facial features from each image. The facial features obtained are compressed to form a vector which is then passed to a Bayesian classifier. This method was tested on three benchmark datasets to show as to how the method overcomes of all the hurdles such as variation of illumination, change of scale, motion of head, change in expression and more. The error rate obtained is in the neighbourhood of 18%.