Finger Knuckle Identification using RLF and Dynamic Time Warping

Abstract

Texture pattern of finger back surface is highly unique consisting of creases and lines which can be used for biometric authentication system. Use of Finger Knuckle Print (FKP) for person identification has been attracted attention of researchers in last few years .Finger Knuckle Print is becoming emerging biometric identifier. In this paper, we present a finger knuckle identification method that uses Dynamic programming (DP) for the alignment of Radon Like Features. The key idea is to use dynamic time warping (DTW) to match Radon Like features of two knuckle images. Experiment is carried out using IIT Delhi finger knuckle database version 1.0. Knuckle features are extracted using the Radon Like Feature technique is classified using DTW for the identification of finger knuckle print. Result obtained using RLF and DTW is promising as compared to subspace and Gabor filtering methods.

Forensic examination of the hand

“…The recognition of this difference in patterning has led to a significant volume of research in the biometric and security‐based literature and the technological methods for capturing and comparing patterns of skin creases across the knuckles. Knuckle creases are universally accepted as a biometric with high levels of discrimination (Choraś & Kozik, 2010; Jaswal, Nigam & Nath 2017; Kumar & Xu 2014; Neware, Mehta & Zadgaonkar 2012; 2014).…”

Feature level fusion using physical biometric traits

“…The outcome acquired was promising and confirms the helpfulness of ROI extraction approach for creating hand-based biometric recognition systems. Neware et al (2015) introduced a finger knuckle identification system that utilises Dynamic Programming (DP) for the arrangement of Radon-like attributes. The main thought is to utilise Dynamic Time Warping (DTW) for matching Radon-like attributes of two knuckle scans.…”