Citation

Munalih, Ahmad Syarif (2014) Secure Multimodal Biometric Fusion using Fingerprint and Finger Vein Descriptors. Masters thesis, Multimedia University.

Full text not available from this repository.

Abstract

Biometrics recognition is a method to verify the identity of a person using a physical or behavioural characteristic. To date, biometrics is widely used as an alternative to password authentication. Fingerprint is of the oldest biometric technique. It uses the patterns of ridge and valley which can be found on the surface of human finger to recognize human identity. Meanwhile, finger vein biometrics is one of the newest biometric technique. Finger vein biometrics uses the pattern of blood vein inside human finger to recognize human identity. Both fingerprint and finger vein biometrics are the main focus of study in this research. A fingerprint feature extraction method, namely Histogram of Oriented Gradient (HOG) is explored and implemented. Besides, an improved finger vein feature extraction method termed as enhanced maximum curvature (EMC) methods is proposed. When EMC is combined with HOG, delicate vein line pattern can be extracted effectively. The proposed method is able to produce better performance as compared to the existing finger vein feature extraction methods. Multimodal biometrics combines more than one biometric sources to address problems such as high intra-class variations, high inter-class similarity and noisy data in unimodal biometrics. In this work, a novel method to combine the HOG features of fingerprint and finger vein biometrics is designed based on Kernel Data Fusion. The method converts the fingerprint and finger vein features into matrix kernels and uses Support Vector Machine for classification. The proposed method has yielded significantly improvement as compared to sole fingerprint or finger vein biometrics as over 99% of recognition accuracy can be achieved with the use of RBF Kernel. Another focus of this thesis is to secure the biometric system to address the revocability and privacy issues. Biohashing is the solution explored in this research.

Item Type:	Thesis (Masters)
Additional Information:	Call No.: TK7882.B56 A36 2014
Uncontrolled Keywords:	Biometric identification
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800-8360 Electronics
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	06 Sep 2017 15:41
Last Modified:	06 Sep 2017 15:41
URII:	http://shdl.mmu.edu.my/id/eprint/6880

Downloads

Edit (login required)