A Comparative Study on Matching Algorithms Used in Biometric Systems and Parameters That Affect Them

Abstract

Biometrics continue to gain popularity due to their ability to uniquely identify an individual as compared to other method. The performance of a biometric system depends on reference threshold. However, the accuracy of the threshold and its adaptability leads to two kinds of errors that these biometric systems experiences namely False Acceptance Rate (FAR) which is a situation where the system accept impostors into the system; and the other one is False Rejection Rate (FRR) which is the number of genuine users who are rejected by the system.

There are many algorithms that have been proposed but their performance and their accuracy remains a major concern in the industry. Due to the nature of biometrics, each feature varies from time to time even though they belong to the same individual.  In this, study we explore and evaluate existing CCN algorithms that have been proposed in addressing the performance of these biometric systems together with parameters that affect them.