Fusion technique for finger knuckle print recognition

This paper presents a novel approach for recognizing finger knuckle print (FKP). In this paper fusion of four techniques (Gabor Feature, MMDA, SIFT-SURF, and Monogenic code) is proposed. Fusion method is used so as to increase the accuracy of biometric systems. Gabor Feature captures the local structure which corresponds to spatial localization, spatial frequency and orientation selectivity. MMDA (multi-manifold discriminant analysis) focuses on graph embedded learning. SIFT-SURF technique educes the local features from a FKP image. Monogenic code is fast feature coding algorithm. The expected outcome of this approach is the recognition rate of the proposed methods, comparison of the methods resulting in best method and comparing the methods on the basis of different file format. Finally these outcomes will be combined for an efficient finger knuckle print recognition.