Biometric Recognition System using Minutiae based Technique Comparing with Correlation Based Technique to Improve the Accuracy

Abstract

Research objective is to improve the accuracy of biometric recognition systems using novel minutiae based technique compared with correlation based technique. Two groups are present in this. Group 1 is novel minutiae technique and second group is correlation based technique each having a sample size of 50. The pretest power is obtained as 80%. There are 100 samples of fingerprints which are predefined in the MATLAB data storage directory named as FVC2002. It is done using MATLAB 2017. The novel minutiae based technique is better in terms of accuracy than the correlation based technique. The false acceptance ratio of novel minutiae based technique is 0 % but the FAR of correlation based technique is 0.01 %. Therefore it is evident, the novel minutiae based technique has a greater advantage over correlation based technique. Significance value p <0.05. 0.8 is taken as G-power value. It is observed that minutiae based technique appears significantly better than the correlation based technique in terms of accuracy and FAR.