Artificial intelligence (AI) has proven its ability to accurately determine whether fingerprints from different fingers belong to the same person. This breakthrough could revolutionize forensic investigations by helping identify if an individual was present at multiple crime scenes.
While existing technologies are capable of matching fingerprints from a single finger, a recent study conducted by Gabe Guo and his colleagues at Columbia University explored the potential similarities between all of a person’s fingertips. They trained a machine learning model using over 50,000 fingerprints from nearly 1,000 individuals. The samples were sourced from public databases, where all fingerprints were either from deceased individuals or had been de-identified.
The trained model was then tested on a separate dataset comprising more than 7,000 fingerprints from around 150 people. The researchers evaluated the model’s accuracy using a statistical measure ranging from 0 to 1. The results indicated that the model scored greater than 0.75, suggesting its reliability in identifying whether prints from different fingers belonged to the same person.
This AI-driven technology has the potential to greatly enhance the efficiency of forensic investigations. Ralph Ristenbatt, a researcher at Pennsylvania State University, remarks on its utility in cases where fingerprints found at multiple crime scenes do not match anyone in existing databases. This technology would allow investigators to determine whether the person who left prints at one crime scene is also connected to prints found at a different location.
Despite these advancements, Guo acknowledges that the current accuracy of the model is insufficient for it to be admissible as evidence in court. He suggests that for legal purposes, the model should be retrained with a larger database to improve its accuracy.
As AI continues to evolve, the application of this technology in forensic investigations holds great potential. With further refinement and expansion of the database, AI could become an invaluable tool for identifying connections between crime scenes and the individuals involved.
The source of the article is from the blog motopaddock.nl