At the moment, all of biometric authentication methods can be divided into two groups: static, when the physical parameters inherent at birth are recognized (fingerprints, iris, retina, facial features, palm vein pattern), and dynamic, when behavioral characteristics are analyzed, such as gait, writing style, etc.
Most of the biometric authentication market is based on static methods, but the share of dynamic and combined methods is growing every year, especially for keyboard handwriting recognition and signature authentication methods.
So far, the most popular and studied technology is the fingerprint one, which is used in criminology and forensic medicine. Each person has a unique whorl pattern that appears at the time of fetus’ skin formation. Whorl patterns are not the same even when it’s twins. And after surface damage, they are easily restored. This is a fast, compact and inexpensive technology. Algorithms recognize Galton’s details — the ends of papillary pattern lines, branches, and intersections. Additionally, information about the morphological structure of the fingerprint (the relative position of closed lines of the papillary pattern, arched and spiral lines) is used. Individual features are converted into a unique code that is stored in the database. The identification process takes less than a second since recognition of each line of the pattern is not required for comparison with samples from the database. It is no longer possible to deceive the device using a silicone dummy with a pattern: capacitive fingerprint scanners register the difference in electrical potentials between the bumps and depressions of the whorl pattern and respond only to the finger of a living person. The method also has some disadvantages: the whorl pattern is easily damaged, and scanners do not read excessively dry or wet skin.
Identification by the iris is also considered statistically reliable due to its genetic uniqueness. The method is based on the formation of up to 266 unique identification points on the corneal image. The recording equipment captures the video image from a distance of no more than a meter, after which a number of actions are performed: selecting the pupil, collecting and counting iris identification points, making a decision on verification after comparison with the database. With age or due to certain diseases, age spots may appear on the iris. To exclude their influence on the recognition result, biometric systems use black-and-white (grayscale) images.
The system can identify 1/3 of the iris if 2/3 of the shell is damaged or shaded, and the probability of error is 1 in 100 000. The system uses infrared lighting to distinguish a live eye from an artificial one or a photo. A significant advantage of this method is its independence from external factors (hairstyle, makeup, beard, etc.) and accuracy. But it's still expensive. Face geometry recognition is one of the three "big biometrics" along with iris recognition and fingerprint technology. There are two-dimensional, which has long been used in criminology (the probability of failure less than 1 %), and three-dimensional, more modern and reliable. The scanning procedure takes about 20-30 seconds. If the face is rotated to the camera, the system compensates for movement and forms a projection of the face with a clear selection of features — the contours of the eyebrows, eyes, nose, lips, etc. — and determine the distance between them. Mostly the template is composed of 12-40 immutable characteristics (depth of eye sockets, shape of the skull, zygomatic arches, the height and width of the cheekbones), due to which the system can recognize a face even with a beard, glasses, scars, hats, etc.
The international Subcommittee for standardization in biometrics (IS0/IEC JTC1/SC37 Biometrics) is developing a unified data format for recognizing human faces based on two - and three-dimensional images. In the future, these two technologies will be combined. Palm vein pattern recognition is a relatively new technology, so its share in the global market is small and is about 3%. The biometric device scans the palm in infrared light and reads its reflection. Hemoglobin in the veins absorbs some of the IR radiation, so the reflection shows a pattern of dark lines — a network of subcutaneous blood vessels. Mathematical algorithms convert the pattern into a digital code and package it into an encrypted template file. To identify a person, the contactless device scans their palm and compares the resulting biometric template with the templates in its database.
In terms of reliability, the technology is comparable to iris recognition: the palm vein pattern is not visible in normal light, so it cannot be photographed or faked, the device scans the palm in several IR spectra at the same time, so it cannot be deceived with a silicone dummy. In the digital template, only some significant parameters of the individual vein pattern are encrypted, so it is impossible to restore a full-fledged image of the hand from it. Even if an attacker gets access to the database, they will not be able to use it in any way. Wherever palm identification is used, this method has proven to be the most convenient and secure among all methods of biometric identity verification.
The BioSmart device, PalmJet, does not need your touch and recognizes the pattern of palm veins even through a medical glove. The reader is indispensable during a pandemic, because it allows you to confirm a person's identity in a non-contact and hygienic way. The device can be useful in all spheres — from IT and network retail to public sector companies. In the future, it can be used for contactless payment of subway fares or, for example, to lock a private house.