By N2H

Factors influencing FAR & FRR in Biometrics

December 18, 2009

FAR in biometrics stands for false acceptance rate, it is the ratio of type 2 errors (when an invalid person is authenticated) to valid authentications. While false rejection rate (FRR) is the ratio of type 1 error (when a valid person is not authenticated) to valid authentications. (Gollmann, 2006)

FAR and FRR are influenced by the following factors:

Sensors being too sensitive causing a valid subject or person not to be authenticated, the sensors can also be insensitive thus causing an invalid subject or person to be authenticated.

Poor picture quality caused by poor lightning in a facial recognition system can also cause problems in biometrics systems thus increasing the rate of FAR and FRR errors.

Another influencing factor in FAR and FRR can be worn out fingerprints that can cause poor quality templates to be obtained during enrollment, this can increase the FAR and FRR.

Noisy surroundings especially in voice recognition can also affect the enrollment process and increase changes of errors in biometrics systems.

The equal error rate (EER) is when the FRR and FAR are equal or close to equal. It is used to measure the performance and accurate of the biometric device. A tradeoff between FAR and FRR can be reached by having a lower FAR and a higher FRR or vice versa. The device with sensitivity closest to the ERR is consider more accurate. Shown below is a Graph of FAR and FRR errors showing the ERR point