Given images of latent fingerprints, identify the closest matching fingerprint in a reference collection of fingerprints. The best algorithms will correctly and robustly identify latent fingerprints across a wide range of image quality.
This is a Reduction-to-Practice Challenge that requires a written proposal and delivery of source code implementing the algorithm. A real-time online scoring utility and leaderboard will be available for this Challenge.
This Challenge seeks to identify innovative methods to perform fully automated matching of digitized latent fingerprints against a gallery of known and unknown reference fingerprints, to support law enforcement agencies’ ability to identify suspects from the scene of a crime. Current Automated Fingerprint Identification Systems (AFIS’s) attempt to match a latent fingerprint to an enrolled print in a gallery and provide a list of candidate matches for a trained Latent Print Examiner.
This Challenge includes a real-time online scoring utility for Solvers to easily score their own code for accuracy on a blinded data set and compare performance versus other Solvers.
Award s for this Challenge depend on three performance thresholds:
The Seeker is not obligated to make multiple awards when multiple submissions meet the requirements. However, where two or more awards meet the requirements for the full award, the quality of the proposal combined with the Seeker's validation of the submitted algorithm will determine if a single or split award will be made of the full amount.
As a Reduction-to-Practice Challenge, Solvers must provide a full and complete written description of the proposed algorithm and source code and any supporting information that validates the performance of the proposed solution. During the evaluation period, the Seeker may validate submitted algorithms using additional (off-line) datasets. The award is contingent upon theoretical evaluation and experimental validation of submissions by the Seeker. To receive an award, Solvers must transfer to the Seeker their exclusive Intellectual Property (IP) rights to the solution.
An InnoCentive RTP (Reduction to Practice) Challenge is a prototype that proves an idea, and is similar to an InnoCentive Theoretical Challenge in its high level of detail. However, an RTP requires the Solver to submit a validated solution, either in the form of original data or a physical sample. Also the Seeker is allowed to test the proposed solution.