The initial test (Phase I) has been completed and aggregate results are available here. Phase I was a proof of concept which demonstrated automated feature extraction from latent images, and the ability to match those features against enrolled 10-print backgrounds and produce candidate lists of manageable size.
Phase II testing commences December of 2007 and expands on Phase I through the use of improved software and larger datasets to give refined estimates of capability. Performance will be explored at both 500 and 1000 ppi, with and without supplementary Regions-of-Interest (ROI) markup. Phase II is limited to those participating in Phase I,
ELFT07 Phase I testing is complete and Phase II has been announced.
NEW
[November 13, 2007]
The Phase 2 Validation Dataset is available as a (42MB) ZIP archive here. After unzipping, see the "readme.txt" file for usage instructions.
[September 26, 2007]
Aggregate Phase I results available here Summary of Results from ELFT07 Phase I Testing
[March 22, 2007]
A new version of the NIST-provided image extraction routines are available here which include support for Cygwin/MinGW GCC, Microsoft C compiler (MSVC), and Linux GCC.
Additional information will be posted at this website as it becomes available.
Interested parties should send contact information (name, email, telephone) to: latent@nist.gov
