COMPARATIVE ANALYSIS OF TWO BIOMETRIC ACCESS CONTROL SYSTEMS

Abstract

This paper compared two Biometric Access Control Systems (BACS). The BACSs employed iris pattern for their authentication. Fast Fourier Transform-driven Access Control System (FACS) uses global iris features while Haar Wavelet Transform-driven Access Control System (HACS) uses local iris features for its template generation. Principal Component Analysis (PCA) was employed to select principal components of the extracted features (local and global). Fuzzy clustering was used for classification and Euclidean Distance (ED) as the distance metric. Experimental result showed that it took more time to train the HACS than FACS because of its intrinsic location in the iris images. It was discovered that global features driven Access Control System (FACS) with EER being 7.78 outperformed the local features driven Access Control System (HACS) with EER of 8.05. Though the two systems satisfied the benchmark of 80% for Recognition Accuracy (RA) of Biometric Systems, FACS exhibited RA of 89.87% while HACS achieved a RA of 83.83% when tested on iris images captured with CMITECH DMX-10 Portable USB 5.0 M pixel CCD Iris Camera at automatic convenient eye distances. Performance of global and local features on other biometric recognition systems can be tested and a means of combining the two features for hybridization can also be sought.