Face Recognition using TSF Model and DWT based Multilevel Illumination Normalization

A facial recognition system is a computer application for automatically identifying or verifying a person from a digital image or a video frame. One of the ways to do this is by comparing selected facial features from the image with a facial database. Non-uniform blurring situations may arise due to tilts and rotations in hand-held cameras. Also degradations occur due to changes in illumination, pose, expression, partial occlusions etc. Here Non-uniform blur and variations in Illumination are considered. Current Face recognition systems consider the motion blur as a space invariant feature and uses simple convolution model for approximating motion blur. But in natural imaging, motion blur is non-uniform. So for face recognition in the presence of space varying motion blur, a methodology comprising of arbitrarily-shaped kernels is used. The blurred face is modeled as a convex combination of geometrically transformed instances of the focused gallery face using TSF Model. The probe image is compared with the convex combinations to find the best match. To handle illumination variations illumination normalization using DWT is used for the test image.