December 22, 2021 by Admin
When configuring a biometric access control system that protects a high-security data center, the system’s sensitivity level should be set:
- to a higher false reject rate (FRR).
- to a lower crossover error rate.
- to a higher false acceptance rate (FAR).
- exactly to the crossover error rate.
Biometric access control systems are not infallible. When tuning the solution, one has to adjust the sensitivity level to give preference either to false reject rate (type I error rate) where the system will be more prone to err denying access to a valid user or erring and allowing access to an invalid user. As the sensitivity of the biometric system is adjusted, these values change inversely. At one point, the two values intersect and are equal. This condition creates the crossover error rate, which is a measure of the system accuracy. In systems where the possibility of false rejects is a problem, it may be necessary’ to reduce sensitivity and thereby increase the number of false accepts. This is sometimes referred to as equal error rate (EER). In a very sensitive system, it may be desirable to minimize the number of false accepts — the number of unauthorized persons allowed access. To do this, the system is tuned to be more sensitive, which causes the false rejects the number of authorized persons disallowed access to increase.