Performance Measurement in Biometrics

All biometric systems are founded on one of three special types of individual’s characteristics. Genotypic characteristics are those that are describes by the genetic makeup of the human being. Instances of genotypic characteristics are face geometry, hand geometry, and DNA patterns. It is significant to note that genotypic characteristics found between the same twins or clones are very alike and often hard to use as a distinctive characteristic to tell the two apart. Genotype means a genetic structure, or a group sharing it, and phenotype is a related term meaning the actual appearance of a feature through the interaction of genotype, growth, and environment. Genetic penetrance explains the heritability of reasons or the degree to which the features expressed.

No biometric technique will be 100% secure, but when contrasted to a PIN or a password, biometrics may offer a superior security. Biometrics in general holds a set of benefits and shortcomings, as the table below summarizes.

The advantages are more important than the drawbacks primarily because of the first reason, biometrics provides positive identification. The fundamental objective is to be able to attain positive identification without having any uncertainties. Since one can’t misplace, forget, or share their biometric information, then it is known positively that the important information cannot be falsified. While it is very complicated to fabricate a biometric characteristic of an authorized user, biometrics (e.g. a face or fingerprint) are not essentially kept a secret.

Performance evaluation in general—and technology evaluations specifically—have been influential in advancing biometric technology. The use of quality detection algorithms in biometrics and propose detection of error trade-off and error versus reject characteristics as measures for the relative assessment of sample quality measurement algorithms. Regardless of the pain taken by international biometric community, the measurement of the precision of a biometric system is far from being completely explored and, ultimately, standardized.

The result of all these discrepancies in measurement (which is minute in the majority of cases) is that each time a template is created from a live biometric characteristic, the consequence is slightly different. Consequently, the result generator is required to make available a matching service to try to establish if the live template belongs to the same human being as the presently chosen master template. The false accept rate (FAR) and false reject rate (FRR) are used to measure if the biometric system is reliable.