Brief by Shorts91 Newsdesk / 03:34pm on 14 Jan 2024,Sunday Tech Today
Columbia University's study challenges the uniqueness of fingerprints, revealing an AI model's ability to determine with 75-90% accuracy if prints belong to the same person. Conducted by an undergraduate student, Gabe Gua, and co-authors, the study employed a deep contrastive network to analyze 60,000 fingerprints. It discovered significant similarities in fingerprints from different fingers on the same person, upending the conventional belief in fingerprint distinctiveness. The AI tool focused on ridges' direction rather than minutiae, presenting potential applications in cold cases but requiring further comprehensive research for real forensic use.
