As AI becomes embedded in forensic workflows, standards for training data quality, algorithm transparency, and output reliability grow increasingly vital. New open-source software and data from the National Institute of Standards and Technology (NIST) are working to connect the expertise of fingerprint examiners with the capabilities of AI—at a moment when technology is reshaping crime scene investigations.
Putting Better Training Tools in the Hands of Examiners and AI Alike
NIST’s collection contains thousands of anonymous fingerprints, gathered in a lab environment from 200 volunteers, who consented to their use for research purposes. Together with NIST’s new open-source software package capable of rapidly assessing print quality, the tools offer a resource to sharpen the skills of forensic scientists and their AI tools.
“These images are good for classroom education, to teach examiners how to look for identifying features,” said NIST computer scientist Greg Fiumara. “And they will also help teach AI algorithms where to look and how to weigh a feature’s importance. With this kind of training, a fingerprint evaluation algorithm will get better.”
The Standards Backing Forensics
A broader framework of standards has long supported the collection, analysis, and dissemination of forensic evidence.
The American Academy of Forensic Sciences (AAFS) has developed many American National Standards to support the forensic, criminal justice, and legal communities. Among them: recommendations for the detection and collection of footwear and tire impression evidence; recommendations for facial approximation in forensic anthropology; and guidelines for the preservation and examination of charred documents.
ASTM International has also developed many American National Standards guiding forensics, with areas of focus including receiving, documenting, storing, and retrieving evidence in a forensic service provider; forensic paint analysis and comparison; and forensic analysis of fibers by infrared spectroscopy.
International standards play a critical role, as well. One example: An ISO/IEC standard on biometric data interchange formats guides the exchange of DNA data used for biometric verification and identification of individuals – supporting compatibility between systems. The U.S. played a leading role in the development of this standard, with ANSI serving as the committee secretariat. INCITS is the administrator of the U.S. Technical Advisory Group (TAG) to this committee.
The convergence of open data, transparent software, and robust standards marks a critical step as AI takes on a growing role in how crime scenes are investigated. Read more about how NIST Helps Fingerprint Examiners With New Data and Software Release.