High Resolution Palmprint Recognition System Using Multiple Features

Abstract

During the past decade, more efforts have been put to utilize palmprint as a biometric identity. This research work presents novel algorithms to extract multiple palmprint features from high-resolution palmprint. The major contributions in this research include the following: 1) Use of preprocessing techniques such as Binarization, ROI Extraction, Enhancement and Thinning. 2)Extraction of multiple features from ridges such as ridge orientation, its frequency, minutiae points (termination and bifurcation points) and pores which during the matching process produces higher accuracy 3) Use of a novel framework in fusion method for an authentication application which performs better than existing techniques. The results of this research work have been compared with the various existing methods and different databases. The comparison results are illustrated through the time taken for execution, ROC and DET curves.