A new AI-powered study from Columbia Engineering questions a long-held belief in forensics—that fingerprints from different fingers of the same person are completely unique. The research found surprising similarities between them.
Led by undergrad Gabe Guo, the team used a deep contrastive network on a U.S. government database of 60,000 prints. Without using traditional fingerprint features, the AI model was able to link prints from the same person with 77% accuracy.
Although forensic journals were hesitant at first, the study was later published in Science Advances. It could lead to better tools for investigations and challenge decades of fingerprint-based identification methods.
#AIinForensics #FingerprintResearch #ColumbiaEngineering #ScienceAdvances #ForensicScience #BiometricTech #AIResearch #DeepLearning #CriminalJusticeTech
Led by undergrad Gabe Guo, the team used a deep contrastive network on a U.S. government database of 60,000 prints. Without using traditional fingerprint features, the AI model was able to link prints from the same person with 77% accuracy.
Although forensic journals were hesitant at first, the study was later published in Science Advances. It could lead to better tools for investigations and challenge decades of fingerprint-based identification methods.
#AIinForensics #FingerprintResearch #ColumbiaEngineering #ScienceAdvances #ForensicScience #BiometricTech #AIResearch #DeepLearning #CriminalJusticeTech
❤1👍1