Fatima Zohra Allam, Latifa Hamami-Mitiche, Hicham Bousbia-Salah,
Volume 33, Issue 1 (IJIEPR 2022)
Abstract
For several years, considerable efforts have been made in the field of biometric research. The major interest of this line of research is linked, among other things, to the recognition of the individual because the security needs are becoming increasingly important, and the economic stakes are colossal. There are many and diverse biometric applications that provide a substantial level of security.
Unimodal biometric systems allow a person to be recognized using a single biometric modality, but cannot guarantee correct identification with certainty. While multimodal biometric systems, using several biometric modalities, guarantee better recognition.
In this article, we are interested in the study of evaluation tools for biometric systems. For this, we will first calculate three essential parameters, namely: False Rejection Rate (FRR), False Acceptance Rate (FAR) and Equal Error Rate (EER). Second, we will determine the performance curves, in this case, the ROC curve (Receiver Operating Characteristic) and the DET curve (Detection Error Tradeoff). The calculation of these metrics allows the evaluation of unimodal and bimodal biometric systems to compare the benefit of merging the biometric modalities.