AI System Can Match Fingerprints to Same Person with High Accuracy, Challenging Uniqueness of Prints
-
Scientists trained an AI model called a deep contrastive network to analyze 60,000 fingerprints to see which ones belonged to the same person.
-
The AI system could match fingerprints to the same person with 75-90% accuracy, challenging the idea fingerprints are unique.
-
The AI focuses on the ridges in the center of fingerprints rather than minute details on the edges.
-
The technique could generate new leads by matching crime scene prints to different fingers of suspects.
-
It could help exonerate innocent people by showing crime scene prints don't match their fingers.