Do prints from two different fingers belong to the same person? AI can tell

Using artificial intelligence (AI), a computer can tell whether fingerprints made by two different digits—say, a thumb and an index finger—belong to the same or different people, a team of roboticists claim. The advance reveals a surprising similarity between an individual’s fingers, but experts question the utility of the observation for forensic investigations. The practice of using fingerprints to identify people took off in the late 19th century and remained the gold standard in biometric identification until DNA analysis emerged. Called dermatoglyphics, it relies on the assumption that the tiny ridges on a person’s fingertips form permanent, unvarying patterns that are unique to that individual. But are the prints on different fingers of the same person similar? At first blush, it might seem like they should be different, too—that’s why you can open your iPhone with one finger, but not another. To find out, Hod Lipson, a roboticist at Columbia University, and colleagues turned to AI. The researchers trained a neural network—a program that mimics the circuitry of the human brain and can learn to recognize and distinguish patterns—first on half a million synthetic images of fingerprints and then on 53,315 fingerprints from 927 individuals. The researchers fed the program trios of images: an “anchor” print from one individual, a “positive” print from a different finger of the same individual, and a “negative” print from another ...
Source: Science of Aging Knowledge Environment - Category: Geriatrics Source Type: research