Are fingerprints unique? According to a study based on artificial intelligence, not really

(CNN) — “Do you think each fingerprint is really unique?”

That’s the question a professor asked Gabe Guo during a casual conversation while he was stuck at home during the COVID-19 lockdown and waiting to start his freshman year at Columbia University. “I had no idea that that conversation would determine the direction of my life for the next three years,” Guo says.

Guo, now a senior in Columbia’s computer science department, led a team that conducted a study on the topic, with University at Buffalo professor Wenyao Xu as one of her co-authors. The study, published this week in the journal Science Advances, appears to call into question a long-accepted truth about fingerprints: Not all of them are unique, according to Guo and his colleagues.

In fact, several journals rejected the paper before the team appealed and it was accepted by Science Advances. “At first there was a lot of resistance from the forensic community,” recalls Guo, who had no experience in the field before the study.

“In the first version or two of our article, they said it’s well known that no two fingerprints are alike. I think that helped a lot to improve our study, because we could then get more data.” Kept recording (and increasing the accuracy), eventually, the evidence was indisputable,” he explains.

A new perspective on scars

To arrive at their surprising results, the team used an artificial intelligence model called Deep Contrastive Networks, which is commonly used for tasks like facial recognition. Researchers weighed in on this, giving it the US government’s database of 60,000 paired fingerprints, which were sometimes from the same person (but from different fingers) and sometimes from different people. .

When working, the AI-based system found that there were strong similarities in fingerprints from different fingers of the same person and was therefore able to tell when the prints belonged to the same person and when not, with accuracy down to the same pair. . which peaked at 77%, which seems to contradict the idea that each fingerprint is “unique”.

“We found a rigorous explanation for why this is: the angle and curvature at the center of the fingerprint,” Guo said.

He further said, over hundreds of years of forensic analysis, people have observed various features called “trait points”, branching and ending points on fingerprint ridges that are used as traditional markers for fingerprint identification. “They are excellent for matching fingerprints, but they are not reliable for finding relationships between fingerprints of the same person,” Guo explains. “And that was the idea that we had.”

The authors say they are aware of potential biases in the data. Although they believe the AI ​​system works very similarly across genders and races, to make this system usable in real forensics, more careful verification by analyzing a larger fingerprint database is needed, according to the study. Is required.

However, Guo is confident that the discovery could improve criminal investigations:

He said, “The most immediate application is that it can help generate new leads for unsolved cases where fingerprints left at a crime scene belong to fingers other than those on file.” “But on the other hand, this will not only help in catching more criminals. It will also help innocent people who will not have to be investigated unnecessarily. And I think that is a win for society.”

Hyperbola?

Using deep learning techniques on fingerprint images is an interesting topic, according to Christophe Champod, professor of forensic science at the School of Criminal Justice at the University of Lausanne in Switzerland. However, Champod, who was not involved in the study, said he did not think the work revealed anything new.

He said, “Their argument that these shapes are somehow correlated between fingers has been known since the beginning of fingerprinting, when it was done manually, and has been documented for years.” “I think they overestimated their article due to lack of knowledge, in my opinion. I’m glad they rediscovered something known, but in short, it’s an exaggeration.”

In response, Guo said that no one has ever quantified or systematically used the similarity between fingerprints of different fingers of the same person, as the new study has done.

“We are the first to clearly show that the similarity is due to the orientation of the ridge in the center of the fingerprint,” Guo said. “In addition, we are the first to attempt to match fingerprints from different fingers of the same person, at least with an automated system.”

According to the authors, the system used in the study to identify similarities between fingerprints could be useful in analyzing crime scenes. Credit: Gabe Guo/Columbia Engineering

Simon Cole, professor in the Department of Criminology, Law and Society at the University of California, Irvine, agreed that the work is interesting, but said that its practical usefulness is overstated. Cole was also not involved in the study.

“We were not ‘wrong’ about the fingerprints,” he said, referring to forensic experts. “This is the unproven but intuitively true claim that no two fingerprints are ‘exactly alike’ “.

According to the article, the system could be useful at crime scenes where the fingerprints found are from fingers other than those on police records, but Cole says this may only happen in exceptional cases, because when the fingerprints are When marks are taken, all 10 fingers and often the palms of the hands are routinely recorded. He said, “It’s not clear to me when they think that law enforcement will have only some of a person’s fingerprints on file, but not all of them.”

The team conducting the study is confident in the results and has open-sourced the AI ​​code so others can test it, a decision that both Champod and Cole praised. But Guo says the significance of the study goes beyond fingerprints.

“It’s not just about forensics, but about artificial intelligence. Humans have been looking at fingerprints since they came into existence, but no one had noticed this similarity until we created our A.I. “This shows the power of AI to automatically identify and extract relevant features,” he says.

“I think this study is the first domino in a huge sequence of these things. We’ll see people use AI to find things that were literally before our eyes, like our fingers, hidden in plain sight “