[ 1 ] [ 2 ] [ 3 ] [ 4 ] [ 5 ] [ 6 ] [ 7 ] [ 8 ] [ 9 ] [ 10 ] [ 11 ] [ 12 ] [ 13 ] [ 14 ] [ 15 ] [ 16 ] [ 17 ] [ 18 ]

Journal of Information Science and Engineering, Vol. 36 No. 2, pp. 337-345

A Proposed Hybrid Biometric Technique for Patterns Distinguishing

Department of Electronic and Computer Engineering
College of Computer Science and Information Technology
University of Anbar
Ramadi, 31001 Iraq
College of Business Informatics
University of Information Technology and Communications
Baghdad, 00964 Iraq
E-mail: maymoonat@gmail.com; maha_mahmood@computer-college.org; raghad.hav@uoitc.edu.iq​

In recent security systems, biometric pattern recognition developed as a major research area. It is of high importance in the process of authentication regarding virtual reality as well as real world entities for the purpose of allowing the system to create an informed decision regarding offering specialized services or allowing access privileges. Recently, the field of security has been a major focus area. The requirement for accurate authentication related to individuals is considered as a main issue in the security field. Old-style approaches of setting up the identity of individual including identification cards, passwords and keys, however, these ways for representing the identity could be easily stolen, lost, manipulated or shared, thus causing security damage. Biometrics traits including voice/face verification, signatures and fingerprints offer a trustworthy choice for identifying or verifying identities and better user acceptability rate. There are two major classes used to represent biometric characteristics; the first is Physiological type that is associated to the body shape such as iris and face recognition, and fingerprints, the other is the behavioural type that is associated to the individual's behaviour such as voice, signature and gait. The paper aims to identify a person by using different multi biometric traits with different technique. This paper handles the two through presenting novel method to the biometric pattern recognition, depending on training neural network (NN) and implements efficient features extraction approach based on SVD, wavelet energy and PCA. Fourth step is the fusion, in this step the three vectors of features that we obtained from the previous step we collect in one vector. The fifth step is testing. In the sixth step, we will compare the database of feature with database this step is called matching and see if the person is existing or not. The quality and accuracy of the identification and recognition of the person are measured in this system by computing the Peak Signal to Noise Ratio (PSNR) and the Mean Square Error (MSE) for face, fingerprint and signature images. The recognition rate of the system is more than ninety. the desired goal of recognizing and identifying a person through his fingerprints, face and signature images. It also shows how the system managed to provide the highest ratio of recognition.

Keywords: artificial neural network, automated teller machine, singular value decomposition

  Retrieve PDF document (JISE_202002_12.pdf)