New Quality Track in NIST FRIF TE

[Fiumara, Gregory P. (Fed)]

Hello NIST friction ridge evaluation stakeholders,

Image quality remains a critical issue in fingerprint identification. ISO/IEC JTC 1/SC 37 has developed a standard for finger image quality, ISO/IEC 29794-4:2024, backed by the reference implementation NFIQ 2 v2.3. SC 37 is actively working to revise ISO/IEC 29794-4, as well as to create a new standard ISO/IEC 29794-12, Fingermark (i.e., latent fingerprint) Image Quality. These revisions seek to potentially include third-party unified quality score algorithms.

To support this standards development process, NIST is adding a Quality track to the FRIF TE program to evaluate such algorithms. FRIF TE Quality implementations must, at a minimum, assign a score [0,100] as to the utility of a fingerprint or fingermark image in a recognition system. Algorithms will be assessed by computing Error vs. Discard Characteristics using information from existing recognition algorithms submitted to NIST. Algorithms can optionally identify image characteristics that help specify quality (e.g., heavy distortion, noisy substrate).

Organizations interested in having their quality assessment algorithms evaluated by NIST, especially those who are interested in contributing their implementation to the international standards community, should thoroughly review the FRIF TE Quality draft application programming interface specification and begin implementation to identify any issues. Questions and comments on the API can be addressed to fri...@nist.gov. Instructions on how to submit an implementation for evaluation will be forthcoming.

Thanks for your continued interest in NIST friction ridge evaluations!

-Greg (@nist.gov)