Biometric template

What is a biometric template

A biometric template is a digital representation of the unique physiological or behavioral characteristics of a person, created as a result of processing biometric data (a face image, fingerprint, voice, vein pattern, iris, etc.). The template is stored in the form of a mathematical model — a set of numbers or vectors that cannot be converted back into the original image or recording. Unlike raw biometric data (photographs, audio recordings, fingerprint scans), the template does not contain personal information that could be used to reconstruct a person's appearance or voice. This makes template storage safer in terms of 152-FZ and requirements for the protection of personal data. Biometric templates are stored in protected databases integrated with cryptographic protection tools (CIPF), and are used in the Unified Biometric System (UBS), access control systems, mobile devices for unlocking, banking applications for confirming operations and in government identification systems.

How a biometric template is created

The process of creating a biometric template includes several stages, each of which uses complex mathematical algorithms and machine learning methods. The first stage is data collection. The device (camera, fingerprint scanner, microphone, IR scanner) records a biometric sample. The second stage is feature extraction. The system identifies key points and characteristics: for a face — the distance between the eyes, the shape of the cheekbones, the depth of the eye sockets; for a fingerprint — characteristic points (minutiae); for a voice — frequency parameters, formants, rhythm, timbre; for veins — points of branching and intersection of vessels. The third stage is transformation into a code. The extracted features are transformed into a numerical vector (mathematical model) — this is the biometric template. The size of the template depends on the type of biometrics: from 500 bytes to 2 KB for a fingerprint, from 2 to 5 KB for a face, from 2 to 10 KB for a voice. The fourth stage is encryption and storage. The template is encrypted using CIPF and saved in a database. It is important that the original biometric sample is not stored if it is not required by the system.

Types of biometric templates

Biometric templates are classified by the type of biometric features used. Physiological templates are based on the unchanging physical characteristics of a person: fingerprints (papillary patterns), face geometry (3D model), vein pattern, iris, retina, hand geometry, DNA profile. Behavioral templates are based on unique behavioral characteristics: voice, signature, gait, keyboard handwriting, mouse manipulation. Multimodal templates combine several biometric features to increase the accuracy and reliability of identification. Multimodal systems provide higher accuracy and resistance to attacks, but require more computing resources and storage space. Standardization of biometric template formats is carried out in accordance with the international standard ISO/IEC 19794.

Advantages of using biometric templates

Using biometric templates instead of raw biometric data gives a number of significant advantages. Security — it is impossible to reconstruct the original data from a template. Efficiency — template comparison is significantly faster, which allows identification in real time (less than 1 second). Standardization — templates are unified for different systems (ISO/IEC 19794). Confidentiality — the template is not considered personal data in some jurisdictions. Memory savings — templates take up significantly less space than raw data.

Protection of biometric templates

Protection of biometric templates is a critically important task, since unlike passwords, biometric characteristics cannot be changed when compromised. Encryption — all templates are stored in encrypted form using certified CIPF and GOST algorithms. Hashing — cryptographic hashing is applied to templates. Salt — a random value is added to the template before hashing. Biometric encryption — the template is linked to a cryptographic key. Distributed storage — the template can be divided into several parts and stored in different protected storages. Cancelable biometrics — a technology in which many different templates for different systems are created from the original biometrics. The design of a system for storing and processing biometric templates requires a professional approach. The design service will help determine the optimal architecture.