Original biometric images are the raw captures taken from a face, fingerprint, or eye. Biometric templates are mathematical representations derived from those captures and used for matching during authentication. Templates are what most systems retain for verification, which means security teams should focus on protecting the template store and limiting unnecessary retention of source images.
How biometric templates differ from source biometric images
The practical difference is that an original biometric image is the raw capture, while a template is the derived representation used for comparison. That distinction matters because the image is usually more sensitive, but the template is the asset most systems depend on for authentication and verification decisions.
It also changes what teams should protect. A compromise of stored images can expose more of the source biometric data, while a compromise of templates can still undermine matching systems and may be difficult to remediate because biometrics cannot simply be reissued like a password.
For a broader view of how biometric factors are used in authentication and verification, the Biometric Authentication and Verification Guide is the best starting point.
Why templates are usually retained instead of raw images
Most production systems retain templates because they are smaller, faster to compare, and often sufficient for matching. That reduces storage overhead and can limit some privacy exposure, but it does not make the data low risk. A template is still biometric-derived data tied to a person, and in some contexts it can be sensitive personal data in its own right.
Retention choices should be driven by necessity. If the system does not need source images after enrollment or quality checks, keeping them creates extra exposure without clear operational benefit. If images must be retained for fraud review, quality assurance, or dispute handling, access should be tightly limited and retention periods should be explicit.
For organisations handling biometric data in the EU, the EU General Data Protection Regulation (GDPR) matters because biometric data can fall into a special category, and the principles of data minimisation and storage limitation become directly relevant to whether source images should be kept at all.
What security teams should protect first
Protect the template store first, then decide whether source images are needed at all. Templates are the material used during authentication, so they become a high-value target for theft, replay-like abuse, model extraction style misuse in some systems, or tampering that degrades match quality. Raw images can also reveal more about the person than the template does, which raises the impact of any leak.
Good controls include strict access control, encryption at rest, audit logging, environment separation, and tightly governed retention. If template and image stores are both present, they should not share broad administrative access, because one compromise should not automatically expose both forms of biometric data.
For the underlying storage and handling discipline, NIST SP 800-53 Rev 5 Security and Privacy Controls provides the control families teams typically use to bound access, logging, configuration, and data protection. Where biometric data crosses cloud boundaries, EU NIS2 Directive is also relevant because it pushes organisations toward stronger ICT risk management, including access control and supply-chain-aware protection of sensitive systems.
Risk and Threat Considerations
Biometric systems are attractive targets because the data is persistent and hard to replace. If an attacker obtains a template store, the organisation may face account takeover, impersonation attempts, or large-scale trust degradation even if the raw images were never exposed. If raw images are retained longer than necessary, the compromise impact can expand from authentication abuse to broader privacy harm.
Failure mechanism: Weak segregation, overbroad access, or poor retention discipline lets an attacker or insider reach the biometric data most useful for abuse, while also increasing the blast radius of any storage compromise.
Impact: The result can be permanent identity exposure, degraded authentication trust, regulatory pressure, and costly re-enrolment or system redesign if the biometric store is no longer considered reliable.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
NIST SP 800-53 Rev 5 sets the technical controls, while GDPR defines the regulatory obligations.
| Framework | Control / Reference | Relevance |
|---|---|---|
| GDPR | Art. 5 — Principles relating to processing of personal data | Biometric image retention is governed by data minimisation and storage limitation. |
| Art. 9 — Processing of special categories of personal data | Biometric data can be special-category data when used for unique identification. | |
| Recommendation — Minimise retention of source biometric images and document the lawful purpose for keeping any biometric data. Apply stricter handling controls and legal basis checks before collecting or retaining biometric data. | ||
| NIST SP 800-53 Rev 5 | IA-5 — Authenticator Management | Templates function as authentication material and must be protected through lifecycle controls. |
| AC-6 — Least Privilege | Biometric stores should only be reachable by narrowly authorised administrators and services. | |
| AU-2 — Event Logging | Access to biometric stores needs auditability for misuse detection and investigation. | |
| Recommendation — Restrict, protect, and rotate authentication material associated with biometric verification systems. Limit access to biometric templates and images to the smallest set of authorised roles. Log all access to biometric data repositories and review logs for anomalous access patterns. | ||
Practitioner Guidance
What to verify: Confirm whether the system actually needs to keep source images after enrollment. If the answer is no, delete or quarantine them under a documented retention rule; if the answer is yes, define the business purpose and the retention window explicitly.
What good looks like: Templates are stored separately from raw images, access is narrowly scoped, and the organisation can explain why each data type exists, who can reach it, and how long it is retained.
Decision rule: If a biometric dataset can authenticate users, treat it as a high-value identity asset, not as ordinary profile data. If you cannot justify continued storage of raw images, the safer default is to stop retaining them.
Practitioner takeaway: The real security question is not whether biometrics are “stored,” but which form is retained, who can access it, and whether keeping the raw image still serves a defensible purpose.
Related resources from NHI Mgmt Group
- Why do biometric systems rely on templates instead of storing full fingerprint or face images?
- What breaks when biometric templates are stored and searchable in a central database?
- What breaks when fingerprint templates are stolen or biometric data is poorly protected?
- Why do centrally stored biometric templates increase identity risk in citizen identity systems?
Deepen Your Knowledge
Free weekly newsletter
Subscribe to the NHI & AI Identity Journal
The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.
Bonus 33% off our NHI Course when you subscribe.
Reviewed and updated by the NHIMG editorial team on November 3, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org