Systems

Overview

  1. Register* (indicating if you wish to be an anonymous participant).
  2. Sign and submit Database License Agreement (you can find it in the downloads page).
  3. Follow instructions for downloading the recipe and methods for spoofing techniques.
  4. Follow instructions for shipping hardware/software systems.
  5. The LivDet IV team will evaluate the systems.
  6. Performance will be assessed.
  7. Results presented at ICB 2015.
  8. LivDet IV System output liveness scores made available to researchers after competition.

* Due to technical problems the account created will be enabled by the administrator as soon as possible.

Submission

Each submitted system must meet the following parameters:

  • Spoof recipe and methods will be provided upon registration for methods. These will be used for testing the system. Two additional unspecified methods will also be used in testing.
  • Participant will ship fingerprint hardware and software system with installation.
  • Installation Requirements:
    • Windows XP
    • 32-bit
    • .exe or similar executable
    • USB or FireWire
  • System output:
    • Collected image*
    • The output is a posterior probability of the live class given the image or a degree of liveness normalized in the range 0 and 100 (100 is the maximum degree of liveness, 0 means that the image is fake). In the case that the algorithm has not been able to process the image, the correspondent output must be -1000 (failure to enroll).

*The collected image will be used in further processing with both quality algorithms and matching algorithms in order to demonstrate fairness between systems as much as is feasible.

System Performance Evaluation

  1. Laboratory staff will systematically attempt to spoof the system and also collect corresponding live data:
    • 1000 live attempts from different people.
    • 1000 attempts for five different materials (three known methods: Play-Doh, gelatin, silicon; two unspecified methods).
    • Same finger molds will be used for all systems. New casts will be made for each system where degradation in spoof sample is seen after multiple placements (e.g. Play-Doh, gelatin).
  2. The parameters adopted for the performance evaluation will be the following:
    Evaluation per system
    • Frej_n: Rate of failure to enroll for the system
    • Fcorrlive_n: Rate of correctly classified live fingerprints for system
    • Fcorrfake_n: Rate of correctly classified fake fingerprints for system
    • Ferrlive_n: Rate of misclassified live fingerprints for system
    • Ferrfake_n: Rate of misclassified fake fingerprints system
    • Fnon-RR**: Rate of failure to acquire for fake fingerprints for system
  3. Declaration of the winner
    • The winner will be awarded based on minimum overall classification error. Only one winner will be awarded.

**More details on Fake Non-Response Rate:  For a system the ability to not acquire a spoof will be considered correct rejection of a fake and calculated as the Fake Non-Response Rate . The ability to not acquire a live finger will be considered incorrect rejection of a live finger and calculated in the Failure to Enroll Rate. In computing the overall average and overall winner, Fake Non-Response Rate and Failure to Enroll will be considered.

Rates are based on the assumption of a threshold of 50. Quality and match scores will be computed based on the collected images to support efforts to maintain fairness between the submitted systems.