How to Evaluate the Accuracy of a Biometric Modality?

Biometrics is a means of identifying and verifying people identity based on physiological characteristics. The core value of a biometric system depends on its ability to correctly accept or reject an individual identity. In other words, accuracy is the criteria to assess the reliability of a biometric modality. The accuracy of a biometric modality is usually measured in terms of matching errors which consist of false acceptance rate (FAR) and false rejection rate (FRR).

With many biometric modalities available such as fingerprint or iris recognition, it is critical for users to assess their accuracy before implementing

With many biometric modalities available such as fingerprint or iris recognition, it is critical for users to assess their accuracy before implementing

False rejection rate (FRR) is the measure of the likelihood that the biometric security system will incorrectly reject an access attempt by an authorized user. This can be particularly frustrating for an end user who knows that they have enrolled in the system and should be recognized and verified.  Biometric readers have different characteristics and limitations that may affect the levels of FRR’s. A voice recognition system may fail to identify a user with a cold. A fingerprint system may be sensitive to dirt or grease or abrasion on the finger. Thus in order to have a low FRR, biometric device should be as insensitive as possible to peripheral factors and have the ability to capture high quality biometric template to match with subsequent biometric scans.

However, the FRR covers only one side of accuracy, false acceptance rate (FAR) also matters. FAR is the measure of the likelihood that the biometric security system will incorrectly accept an access attempt by an unauthorized user. Like FRR, the FAR depends on several factors such as the ability to capture the biometric information accurately, the ability to match it correctly and more importantly, the uniqueness of the chosen biometrics. In this case, iris recognition is an ideal biometrics where there is a high degree of randomness and complexity in iris patterns that underpins its uniqueness and distinctiveness between individuals. FRR and FAR both play an important role in the evaluation of biometric performance, for a correct and reliable figures, users are recommended to refer the result of a trustworthy organization or institute that conducts evaluation test between biometric providers.

Taking an example, the National Institute of Standards and Technology (NIST) have conducted many evaluation tests regarding fingerprint (Fingerprint Vendor Technology Evaluation – FpVTE test), face (Face Recognition Vendor Test – FRVT), iris (Iris Exchange – IREX test) and multiple biometric (MBE test) which uses FAR, FRR as one of the evaluation metrics. NIST-proven accuracy is considered an international recognition for the company’s superior technology. In India, where UIDAI – the biggest national ID project worldwide was born, technology providers have to be granted STQC (Standardization Testing and Quality Certification) in order to step into its market.

There are many biometric modalities on the market nowadays; therefore, it is critical for users and system integrators to have sufficient information and reliable assessment before they make a purchasing decision. Choosing the proper biometric modality for your project can be better served by researching not only about their price, user friendly, longevity but also the degree of accuracy they could bring.