Indentification

Identification operations are those in which the biometric data itself is used to identify a database entry or digital record relating to the individual. Identification is often the instinctive biometric use case we may assume reflects a majority of usage - fed by Hollywood images of hands-on glowing screens or experience with Law Enforcement use cases involving facial recognition and surveillance tools.

Warning: Identification use cases pose more hazards as the biometric system becomes responsible for an adjudication - or decision-making - step that relies on the matching process. As we will discuss later in this guide in accuracy considerations, this can introduce greater chances of exclusion.

While the underlying technology may be virtually identical (a system deployed to do 1:1 matching may be easy to reuse for 1:N matching through reconfiguration), these nonetheless involve distinct processes and technology - in which the same operation is run multiple times, and decisions are made primarily based on thresholds with no independent check such as the presentation of an ID document.

In practice, there may also be use cases where 1:N matching is more dangerous - allowing a system built for a distribution to be used for surveillance or law enforcement - and opening more latitude for exclusion where the technology fails.

Info: 1: N matching systems with built-in enrolment may be our ‘intuitive’ sense of how biometrics work; for instance, surveillance systems in public places or as part of security applications often use sensors built into cameras, effectively enrolling many subjects simultaneously. However, in practice, due to accuracy, data collection, dignity, and other requirements, many commercial and humanitarian applications can work a little differently depending on the system and use case.