A biometric template is a stored representation of biometric features, such as a fingerprint pattern, face geometry, iris structure, or voice characteristics, that a system uses for matching or verification. It is usually not the raw image or recording itself, but a feature set or model derived from it during enrollment. When a user later authenticates, the new sample is compared against the template to decide whether the match is strong enough.
This matters because templates are sensitive identity data. If attackers steal them from a database or intercept them in transit, they may be able to use them for impersonation, cross-service correlation, or, in some cases, reconstruction of the original biometric. Good defenses include encrypting templates at rest and in transit, storing as little as possible, and using template-protection methods such as cancelable biometrics or secure hardware. Privacy-preserving systems can also verify a person without exposing the full template to the verifier, reducing the impact of a breach.