The essence of biometric security consists in the use of electronic identification systems based on unique biological features and in determining the access rights of this person to one or another system. Biometrics deals with two issues: data acquisition and verification (identification of the personality) – comparison of newly obtained data with data obtained earlier and stored in memory. Identification of fingerprints by today is the most commonly used biotechnology.
The basis of any biometric access control system is a sensor that “removes” some physical characteristics of a person. The output signal of the sensor is translated into a digital form, extracting only useful information using mathematical algorithms and preserving the received information in memory.
When identifying a person, a reprint is taken, transforming it into a template using the same algorithms as when inserting the template into memory and comparing the two templates obtained with each other to decide on their identity.
In systems that work with fingerprints are used two main methods: correlation and method of miniatures. The first one is rather universal and can be used practically all biometric technologies. The second is the main method in fingerprinting, but the comparison of the template in memory and the newly received template is complicated by the fact that it is impossible to attach the finger to the scanner twice equally. In this case, there is not only a shift and rotation of the drawings, but also deformation, the loss of part of the miniatures.
The problem of reduction of hardware expenses can be solved having replaced identification that is comparisons of the received template with the database of earlier remembered reference templates, verification that is comparisons of the print which is again received when scanning with the only sample. This only sample gets out of the database because the user identifies previously also make introduction of a code from the keyboard, presentation of the payment card or other similar way.
Dactyloscopic systems do not guarantee the correct identification, on this account, introduced two very important parameters: the level of false admission and the level of false failure. The first determines the probability of admittance to the third party protected when the system accepts it as its own. The second parameter, on the contrary, determines the probability that a person whose imprint in the database will not be allowed (from the first time, from the second or third time, perhaps, a person may be passed).
And it is one more sad fact… Some people do not have a papillary pattern on their fingers. There is a following statistics: about 5-8% of people have “bad” prints from the point of view of recognition, and from 2 to 3% do not have them at all in general (various pathologies and skin diseases). The facts of making false prints are also established.