Journal of Engineering Science and Technology
Vol. 18, No. 6 (2023) 2915 - 2927
© School of Engineering, Taylor’s University
2915
ENHANCING FINGER OUTER KNUCKLES
RECOGNITION USING DEEP RECURRENT NEURAL NETWORK
RAID RAFI OMAR AL-NIMA
1
, ABDULRAHMAN W. H. AL-ASKARI
1
,
KARAM M. Z. OTHMAN
1,
*, ABDULRAHMAN K. EESEE
1,2
1
Northern Technical University, Mosul, Iraq
2
ELKH-PE Complex Systems Monitoring Research Group, Department of Process
Engineering, University of Pannonia, Veszprem, Hungary
*Corresponding Author: karam.mz@ntu.edu.iq
Abstract
In recent years, Finger Outer Knuckle (FOK) has come out as a promising biometric
modality. This paper considers enhancing the FOK recognition of verification by
using a suggested efficient deep learning network. It is called the Deep Recurrent
Neural Network (DRNN). This network has the ability to deal with both minor and
major FOKs. It consists of input layer, hidden layers, output layer and global
feedback. It can further increase the verification performance. Images of minor and
major FOKs from the Indian Institute of Technology Delhi Finger Knuckle
(IITDFK) dataset are employed. The result demonstrates promising accuracy
verification rate of 96% after utilizing both major and minor FOKs.
Keywords: Biometric, Deep recurrent neural network, Finger outer knuckle,
Recognition.