🆕 API Unification & FpVTE 🎉

[Fiumara, Gregory P. (Fed)]

We're excited to announce two new endeavors today—the return of FpVTE and a unification of our fingerprint technology evaluation APIs.

API Unification

We're thrilled that so many of you participate in all our existing fingerprint technology evaluations—MINEX, PFT, SlapSeg, and ELFT. Each of these evaluations necessarily has its own API representing the core functionality under test. However, much of the supporting infrastructure in the API (e.g., finger position codes, template data, product identification) is reimplemented in each, and in some cases, implemented subtly differently, requiring different implementations of similar code for each evaluation.

The goal of unification is to have a single API for these common infrastructure pieces for all evaluations. At a minimum, this should allow you to share code across submissions to different NIST evaluations. At the extreme end, this could allow you to use a full SlapSeg implementation directly in an ELFT feature extraction implementation, and potentially be able to submit to both evaluations simultaneously.

Documentation for the draft unified API can be viewed online. Since we're unifying the API, we'll need a unified name as well...

Friction Ridge Image and Features

We're calling our unified technology evaluation program Friction Ridge Image and Features (FRIF) Technology Evaluation (TE). You can call it FRIF (rhymes with "cliff") or FRIF TE (rhymes with "thrifty") for short—and we will too, because it's entirely too long of a name. To be clear that our evaluations use the same common API, we're planning to rename our existing evaluations as well:

• ELFT will become Mark One-to-Many Search,
• PFT will become One-to-One Comparison,
• MINEX will become PIV One-to-One Comparison, and
• Slap Fingerprint Segmentation will shockingly maintain its existing name.

There are other important reasons to change the evaluation names—to better represent what the evaluations do, who we accept submissions from, and align with standardized terminology. This also simplifies our communication, maintaining a single contact and web presence for all evaluations.

This unification won't happen overnight, but it will start immediately with...

Exemplar One-to-Many

Formerly known as FpVTE, FRIF TE Exemplar One-to-Many (FRIF E1N or E1N, for short) marks the return of an exemplar friction ridge image one-to-many search technology evaluation. This is an evaluation of algorithms at the core of many automated biometric identification systems (ABIS), colloquially known as "tenprint search." Implementations will be asked to store extracted features from one or more friction ridge images or feature sets into a template and search those templates against implementation-defined reference databases. Analysis like that produced by ELFT will be generated for submissions in an automated fashion. 

Anticipated improvements from FpVTE 2012 include:

• faster speed requirements,
• support for palm images (including potential palm search),
• larger databases, and
• support for feature sets.

NIST will operate E1N in the same ongoing manner as all our other TEs. E1N has been designed from the start to use the FRIF common API, and applies lessons learned from our existing TEs. As a new evaluation, we will target Ubuntu 24.04 LTS as our operating system. Documentation for the draft API and a draft test plan are available now.

Our initial timeline is to release a finalized API and begin receiving submissions for this evaluation in Fall 2024.

What's Next?

If you received this message directly from NIST, it means you're a member of one of our evaluation mailing lists. We're going to attempt to subscribe you to our new mailing list, fr...@list.nist.gov, and you'll receive a notification to that effect. You can otherwise subscribe by sending an email to frif+su...@list.nist.gov. We'll continue to use our existing evaluation mailing lists until those evaluations have migrated to the FRIF common API.

We ask that you please review the FRIF common API, the FRIF E1N API, and the FRIF E1N Test Plan and submit comments and suggestions to fr...@nist.gov (privately) or on GitHub (publicly).

Thank you for your assistance, your patience as we implement this transition together, and most importantly, for your continued participation!

Thank you!

-Greg

-- 
Greg Fiumara
Image Group -- NIST