Multiscale Logarithm Difference Edgemaps for Face Recognition Against Varying Lighting Conditions - 2015
Lambertian model is a classical illumination model consisting of a surface albedo part and a lightweight intensity element. Some previous researches assume that the sunshine intensity part mainly lies in the massive-scale features. They adopt holistic image decompositions to separate it out, but it is troublesome to come to a decision the separating purpose between massive-scale and small-scale features. In this paper, we tend to propose to take a logarithm remodel, which can modification the multiplication of surface albedo and light intensity into an additive model. Then, a difference (substraction) between 2 pixels in a very neighborhood will eliminate most of the light intensity element. By dividing an area into subregions, edgemaps of multiple scales can be obtained. Then, every edgemap is multiplied by a weight that can be determined by an freelance coaching scheme. Finally, all the weighted edgemaps are combined to make a sturdy holistic feature map. Extensive experiments on four benchmark knowledge sets in controlled and uncontrolled lighting conditions show that the proposed method has promising results, especially in uncontrolled lighting conditions, even mixed with alternative complicated variations.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here
- Iterative Vessel Segmentation of Fundus Images - 2015
- Localized Multi-Feature Metric Learning for Image Set Based Face Recognition - 2015