Columbia University has developed an AI tool that can identify fingerprints with 75-90% accuracy, potentially impacting biometrics and forensic science.
The team trained the tool to examine 60,000 fingerprints to determine which ones belonged to the same individual.
Researchers suggest the AI tool analyzes fingerprints differently from traditional methods, prioritizing the orientation of ridges in the finger's center over individual ridge endpoints and forks.
The tool diverges from traditional forensics markers, relying on the curvature and angle of swirls in the center, a departure from decades-old practices.
The results could have the potential to impact both biometrics and forensic science, as unidentified thumb and index fingerprints could be forensically connected to the same person.
However, more research is needed to develop this technology further, as AI tools are typically trained on vast amounts of data. The study has been peer-reviewed and will be published in the journal, Science Advances.
