follow us on facebook follow us on twitter email us

Nail to Nail (N2N) Fingerprint Challenge

About the Challenge
Can you build the best autonomous nail-to-nail fingerprint scanning device?

Posted By: Intelligence Advanced Research Project Activity
Category: Scientific/Engineering
Skill: Engineering Interest: Science & Research Submission Dates: 12 a.m. ET, Feb 06, 2017 - 4 p.m. ET, Mar 17, 2017 Winners Announced: Nov 30, 2017

Registration is now open!  Registration is different than what had been discussed at participants’ day.  Participation is NOT a first come first served approach!  All participants who submit a registration form by the March 17 progress to Stage 2.

The Nail to Nail (N2N) Fingerprint Challenge seeks to identify fingerprint collection solutions able to acquire nail-to-nail image capture of the friction ridge surface without the need of a human operator. Collection of N2N, or rolled, fingerprint images allows for improved recognition performance in live and forensic matching scenarios. Traditionally, fingerprints can be grouped into three types: plain (or slap), rolled (or nail-to-nail), and latent (i.e., those found at a crime scene which must be developed through dusting, fuming, or other techniques). Plain prints represent information from only the center portion of the finger pad, whereas rolled prints represent information around the entirety of the finger pad, from one nail edge all the way around to the other nail edge. Latent prints are those left behind on a surface when the person is no longer present. Latent prints are typically partial or degraded in quality, due to the nature of how they are left behind or imprinted, unwittingly, on an object’s surface.

In traditional matching scenarios where plain or rolled prints are compared against one another, the larger surface area translates into more discriminative information for matching. For forensic applications, the larger surface area in the reference image increases the likelihood of obtaining sufficient overlap when matching partial latent fingerprints. Plain prints are easy to collect and typically don’t require operator assistance to produce good quality images for matching. While N2N prints provide superior information for matching, they are more difficult to collect than plain and require physically rolling the finger across a flat surface, often requiring assistance from a human operator. Having an operator involved in the process constrains the feasibility of collecting rolled prints in a variety of environments and operational scenarios. Removing the human-in-the-loop from this process through advanced collection and processing techniques will allow better fingerprint data to be collected, leading to improved recognition performance, while reducing the time and cost of collection.


The goal of the N2N Fingerprint Challenge is to improve biometric fingerprint collection and recognition systems by eliminating plain fingerprint captures. This challenge seeks to identify solutions which can perform live capture of N2N fingerprints without requiring a human operator1 for the purposes of matching against other latent or live capture of fingerprints. This unassisted N2N collection will allow for more distinguishing data to be collected while also alleviating the time and cost associated with using human operators. The developed system should collect fingerprint data that performs as good as, or better than, existing operator controlled N2N fingerprint collection approaches. Performance of the developed N2N collection systems will be evaluated using data collected from a live test using human subjects and encompasses both live and latent fingerprints. The participant collected data will be compared against “gold standard” N2N and latent data using conventional fingerprint recognition algorithms. Participants will be judged based on traditional biometric performance measures in addition to speed of the collection process. Participants are not required to develop algorithmic/software techniques to match N2N or latent data.

Participants are expected to design a system that will perform live capture of fully cooperative subjects. It is expected that subjects will be in close proximity to the device, as such, contactless or standoff systems are not required, but are still in scope if they meet the requirements for N2N collection. Systems may be facilitated by an observer, but the observer may not physically interact with the subject. Additionally, mechanical arms or other devices which “roll” the finger as a human operator traditionally would are out of scope but other mechanical components may be allowed.

The N2N Challenge is a three-stage process: 1) registration and feasibility review 2) system build and judging, and 3) Live Test: evaluation of performance. Note: All submissions will be accepted through only.

Review additional timeline and challenge resource material on the N2N challenge information site.

Chris Boehnen Ph.D.
Senior Program Manager / IARPA
James A. Loudermilk
Senior Level Technologist / Federal Bureau of Investigation
Arun Vemury
Program Director / DHS Science & Technology Directorate
Judging Criteria

Grand Prize: Best Useable Matching System

The Best Usable Matching System Grand Prize is awarded to the Best Latent Matching system.

- No more than 20% slower than existing approaches
- N2N matching no worse than 2% of legacy/baseline
- Latent matching performance no worse than 2% of legacy/baseline 90% of subject data captured

Gallery Accuracy Prize

Gallery Accuracy Prize is awarded to the system with the Best N2N match performance.

- Must be no slower than existing approaches
- 90% of subject data captured

Latent Accuracy Prize

The Latent Accuracy Prize is awarded to the system with the best Latent match performance.

- Must be no slower than existing approaches
- 90% of subject data captured

Speed Prize

Speed Prize is awarded to the system with the Fastest N2N capture time.

- Latent matching must be within 80% of the N2N baseline method
- 90% of subject data captured

Print Provider Prize

The Print Provider Prize is awarded to up to 12 Live Test participants that provides their captured data to be open source and available to the public.

Master Builder Prize

The Master Builder Prize is awarded to any of 12 invited Live Test Participant that comes to attend the Live Test.

How to Enter

Challenge Stages and Prizes

Stage 1:  Registration and Feasibility Review

Participants will fill out a registration form providing information about their proposed solution. The application will require the following sections:

  • Abstract – brief description of what you plan to build
  • Solution Description
    • Anticipated hardware and software components
    • High Level system diagram
  • Usability – how will the user interact with the device?
  • Innovation – what’s the novelty of this solution, has it been tried before?
  • Safety Assessment – are there any components (electrical components, illuminators, etc.) in your design which may cause safety concerns with human subjects testing?

Applications should not be more than 2 pages and should describe the anticipated solution with as much detail as possible.  The government will provide informal feedback on the application to the applicants.

During future stages of this competition, outlined below, participants will be required to submit additional information.

Stage 2:  Development and Judging

Builders will have until July 2017 to develop their solution for the N2N challenge.  Builders will be asked to submit sample fingerprint imagery from 3 subjects collected from their systems along with a short video of the system as a “proof of build” and certification to attend the Test and Evaluation (T&E) event taking place in September 2017.  Devices will also be judged for safety at this time.  Builders who do not provide certification of intent to attend the T&E event will be automatically disqualified from moving to Stage 3.    Winners may continue to further refine their devices prior to the live test.

Winners will be selected to participate in the Live Test  The videos produced by the builders will be evaluated along with an updated abstract  and sample imagery to determine participants selected for the next stage.

The submissions will be judged by a 3 person USG team based on their content and not the quality of the video production.  The sample imagery may be used to confirm image standard compliance and interoperability with the USG API used for evaluation.  Additionally, the sample imagery will be used to determine a cursory measure of image quality.  The Judges will select the stage 2 winners based upon their assessment of the viability, novelty, and operational usefulness of the technology and the criteria for Stage 2 outlined below.  In the event that more entries are received than space is allocated, prize challenge balance may also be considered to ensure a diversity of approaches.

Stage 3: Test & Evaluation and Prize Awards

The government will conduct a test where human subjects will present finger biometrics to the maker’s developed system.  To support testing, prize participants are responsible for operating their own capture device during the testing, ensuring safety and IRB compliance of the developed system with testing staff, and providing collected data in the specified format described in the N2N Fingerprint Image Format document for evaluation.

In the event that a participant is unable to produce a Print Data Set due to unforeseen circumstances, invited participants may still be eligible to receive a prize for completing Stage 2 on the individual assessment of IARPA.

Testing will include collection of baseline data using current “gold standard” techniques for N2N, using a skilled operator, and latent fingerprints which will be collected by the government team.  Two baseline systems will be used to collect N2N data by the government.  Prize participants will collect N2N data from subjects without an operator. To be eligible to proceed to the evaluation stage and win the awards, makers are required to produce N2N data on 90% (less than 10% Failure to Acquire [FTA]) of the subjects provided without throwing away or filtering unacceptable data.  Participants are not allowed to intentionally throw away data that the system accepted and captured.  Captured data must include prints for all ten fingers to be considered complete.  Please note that prize participants are only required to provide N2N live data and will only be assessed on their N2N submissions.   All captured data must be provided and the amount produced will be determined by the total number of subjects who interacted with the device.  Makers must keep up with the pace of test subjects (approximately 5 minutes per subject to collect all required data) and ensure their collection process does not hinder the overall collection, impacting other participants.

Makers will need to operate their device in the MdTF.  At the facility, each team will be provided with a 3’x6’ table, a 6-plug power strip, and a network cable to hook up to the test facility network.  All equipment for testing must fit on or under the provided table.  Makers will need to configure their device to send images to the facility collection API, which will also be used to calculate time per each participant.

Grand Prize: Best Useable Matching System $100,000.00 The Best Usable Matching System Grand Prize of $100,000 is awarded to the Best Latent Matching system.
Gallery Accuracy Prize $25,000.00 The Gallery Accuracy Prize of $25,000 is awarded to the system with the Best N2N match performance.
Latent Accuracy Prize $25,000.00 The Latent Accuracy Prize of $25,000 is awarded to the system with the best latent match performance.
Speed Prize $25,000.00 The Speed Prize of $25,000 is awarded to the system with the Fastest N2N capture time.
Print Provider Prize $96,000.00 The Print Provider Prize is awarded to any Live Test participant that provides their captured data to be open source and available to the public. Up to 12 prizes of $8,000 will be awarded.
Master Builder Prize $24,000.00 The Master Builder Prize is awarded to any invited Live Test Participants that attends the Live Test. Up to 12 prizes of $2,000 will be awarded.

Add to the Discussion

No solutions have been posted for this challenge yet.
Submit Solution
Challenge Followers
Public Profile: 1
Private Profile: 7