The Oldest Biometric
Fingerprint identification is the most established forensic identification method, used in criminal investigation for over a century. In 1892, Argentine police officer Juan Vucetich made the first criminal fingerprint identification, solving a murder case in Buenos Aires. Today, automated fingerprint identification systems (AFIS) search databases of hundreds of millions of prints in seconds, but the underlying principle remains the same: no two fingerprints are alike.
Minutiae: The Fingerprint Alphabet
Fingerprint comparison centers on minutiae — the points where friction ridges end or bifurcate. A typical fingerprint contains 80 to 120 minutiae, each characterized by its type (ending, bifurcation, or dot), position (x,y coordinates), and ridge direction (angle). The specific configuration of these points creates a pattern complex enough that random duplication is statistically impossible across the human population.
Matching Algorithms
Modern AFIS algorithms work in two phases. First, alignment: the system finds a geometric transformation (rotation, translation, scaling) that best overlaps the candidate minutiae with the reference. Second, scoring: matched minutiae pairs within spatial and angular tolerances are counted and weighted by their rarity. Minutiae in unusual positions contribute more to the match score than those in common locations.
Threshold and Error Rates
Every matching system faces a tradeoff between false positives (declaring a match when prints are from different people) and false negatives (missing a true match). Lowering the match threshold catches more true matches but increases false alarms. This simulation lets you explore how the number of matching minutiae and tolerance settings affect match confidence and error rates.