Before computers, searching a fingerprint collection meant manual classification and hours of filing. Today an Automated Fingerprint Identification System (AFIS) can compare a single mark against tens of millions of records in minutes. But it's widely misunderstood — an AFIS does not, by itself, declare a match.
How it works
When a print is entered, the system:
- encodes it into a digital template — chiefly the map of minutiae (their types, positions and orientations),
- searches that template against the database using matching algorithms,
- returns a ranked list of candidates with similarity scores.
The output is a shortlist of possibilities, not a verdict.
The human stays in the loop
A trained examiner then compares the crime-scene mark against the candidate prints using ACE-V, and makes the actual identification or exclusion. This distinction is critical: the AFIS narrows the field; the examiner decides. Treating a high-ranking AFIS candidate as an automatic match is exactly the error that contributed to the Brandon Mayfield misidentification.
Standards and interoperability
Large-scale systems rely on shared data formats — notably the ANSI/NIST-ITL standard — so that prints captured on one system can be exchanged and searched on another. That interoperability is what lets national and international databases work together, and it's central to any modern fingerprint or biometric enrolment system.