As every industry has its own terms, biometrics industry also has many specific terms. This article introduces some of them.
It stands for Equal Error Rate. This is the error rate calculated by adjusting the threshold from the measurement result to unify FAR and FRR. This is often referred as an indicator for representing the accuracy performance of authentication technology. The lower the value is, the higher the accuracy.
This means a series of processes for creating a template. For finger vein recognition or fingerprint recognition, it is done by placing a finger on the device.
It stands for False Acceptance Rate. It is the rate that one user is authenticated as a different person by mistake.
It stands for False Matching Rate. In the case of finger vein recognition or fingerprint recognition, it is the rate that the matching process for two different fingers is regarded the same by mistake. This indicator means a probability for one matching process.
It stands for False Non Matching Rate. In the case of finger vein recognition or fingerprint recognition, it is the rate that matching process for same fingers is regarded the different. This indicator means a probability for one matching process. For example, on the system that one user can place a finger by three times, if third trial is succeeded after two times failures, FNMR would be about 66.7% since it is regarded as 2 errors out of 3 trials, while FAR would be 0% since it is regarded as 0 error out of 1 whole trial.
It stands for False Recognition Rate. It is the rate that a person fails in authenticating with their own template.
It stands for Failure To Enroll Rate. It is the rate that one user can’t enroll their biometric information due to some reason.
Identification and Verification
Identification is to specify one specific person out of many candidates. It is also called 1 to many authentication or 1:N authentication. In general, the system does matching process with all biometrics data even if one very similar candidate is found.
Verification is to determine if one user matches to one specified biometrics data. It is also called 1 to 1 authentication or 1:1 authentication. Unlike Identification, the system generally does matching process with only one specified data.
This is biometric information enrolled for a specific biometric modality.
This is a value how far differences should be regarded as matched when comparing pre-enrolled data and matching data. By modifying this threshold value, you can adjust the balance between accuracy and usability like ‘To decrease FAR instead of increasing FRR’ or ‘To decrease FRR instead of increasing FAR’.
In general, it means explicit or potential weak point of security on a specific system. One of typical example in biometrics is that offering biological characteristic information made of artificial object. The benchmark against offering an artificial object is called vulnerability assessment, which is used as an indicator that show the performance of biometrics modality as well as error rate evaluation.
Authors of this article
General Manager, Technology Promotion
Senior Sales Manager
General Manager in Marketing