A Novel Approach Based on Fingerprint Identification and Face Recognition

Neha Yadav, Vinod Kumar, Monika


In real world applications unimodal biometrics have some limitations such as noise sensitivity, data quality, non-universality,
interclass difference. These issues can be tackled easily by multimodal biometric system. This paper investigates the fusion of face visual
features by using e-PCA that resolves the issue of visual difference in face classification process. An implementation of face and fingerprint
fusion using PCA based and hamming distance for calculating minutiae features in fingerprint is fused at matching score level .This paper uses
match score level by choosing the id of both face and fingerprint Better result can be obtained using Euler PCA if multiple instances of same
person were used for matching and combining. Templates of fingerprint and face feature are tested on self created image database. PCA
observes different dimensions and evaluate the principle dimensions where the variation is high. This proposed work is tested in MATLAB
2015b with performance metrics FAR of 3.33%, FRR of 3.27 and accuracy by changing the person instance feature at every time.


Keywords: Biometrics, ePCA, PCA, Biometrics Technique, Multimodal Biometric

Full Text:


DOI: https://doi.org/10.26483/ijarcs.v7i3.2689


  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science