Artificial Intelligence Can Identify if Fingerprints from Two Different Fingers belong to the Same Person

Artificial intelligence can accurately identify whether fingerprints left by different fingers belong to the same person. This helps forensic investigators determine whether one person was at separate crime scenes.

Current technology can only match fingerprints left by the same finger. However, previous research suggests that all human fingertips may have fundamental similarities.

So, Gabe Guo Researchers at Columbia University in New York trained a machine learning model to determine whether fingerprints from different fingers can be identified as belonging to the same person. More than 50,000 fingerprints from around 1,000 people were used in the training. Samples were obtained from public databases at the National Institute of Standards and Technology and the University at Buffalo, New York. All fingerprints either belonged to deceased individuals or were anonymized from those living.

The team then tested the trained model on another set of more than 7,000 fingerprints from about 150 people. They evaluated the model using a statistical measure that estimates accuracy on a scale of 0 to 1. The researchers found that the model's score was greater than 0.75. This suggests that the model can reliably identify whether fingerprints from different fingers belong to the same person.

This technology has the potential to improve the efficiency of forensic investigations. “It could be useful if fingerprints found at multiple crime scenes don't match anyone in the database,” he says. ralph listenbutt at Pennsylvania State University. “Is the person who left fingerprints at this particular crime scene the same person who left them?” [different] What about this other crime scene print? ”

However, “the accuracy is not sufficient at this time.” [for this model] The court will have to decide this,” Guo said.

“If this is actually used for legal purposes, it will require professional retraining. [bigger] database” Hod Lipsonalso part of the research team at Columbia University.