Modified Scale Invariant Feature Transform for Finger Knuckle Print Recognition and Authentication

Umar Abdullahi,Rabi Mustapha,Khalid Haruna,and Aminu Usman Jibril

Keywords: Biometric Image, knuckle print, blur image, blobs detection, invariant image


The time cost of Finger Knuckle Print (FKP) systems using Scale Invariant Feature Transform (SIFT) algorithm is one of the major issuesthat directly affect the efficiency and operations of FKP based recognition and authentication system using SIFT. In this research study SIFT was enhanced with the capability to recognize and authenticate FKP at a significantly reduced time. This in order to demonstrate the time efficiency and authentication ability of the enhanced SIFT, experiments were conducted using challenging benchmark datasets of the Finger Knuckle Prints which are the PolyU datasets. The experimental results show that the average time for FKP recognition and authentication is about 0.08s and yielded significant recognition capability in detecting FKP of low resolution found in the datasets