Illumination Normalization Using Logarithm Transforms for Face Authentication

Abstract

In this paper we propose an algorithm that can easily be implemented on small form factor devices to perform illumination normalization in face images captured under various lighting conditions for face verification. We show that Logarithm transformations on images suffering from significant illumination variation, produce face images that are improved substantially for performing face authentication. We present illumination normalized images from the CMU PIE database to demonstrate the improvement using this non-linear preprocessing approach. We show that we get improved face verification performance using this scheme when training on frontal illuminated faces images, and testing on images captured under variable illumination