Updated Validation and Datasets

[Evaluation of Latent Friction Ridge Technology (ELFT)]

Hello ELFT Stakeholders, 

Validation
We have updated the validation code and images to try and catch several types of issues we've seen over the past few months. If you previously requested the image dataset from our website, you will receive a link to download the updated dataset shortly. If not, you will need to request it from our website. You must use the new code and images for any submissions received after today.

Data
We are working to bring a new larger operational dataset to the test soon. The scenario is still searching a single latent finger/palm against a plain/rolled/palm exemplar database. Timeline for rollout is still being determined. Barring issues, all existing participants will have results for this new dataset posted when ready. We may choose to remove some of our less statistically-significant datasets as a result.

Scenarios
We announced a limited scope to start the test in April 2021, to help participants ease into the evaluation. We have since had several organizations participate successfully and plan to broaden the test scope in the coming months. In particular, we believe we have acquired enough data to perform "reverse latent" searches (i.e., latents in the database), and this will likely be the added following the aforementioned new dataset. As a reminder, all scenarios are tested at least once in the ELFT validation package (and have been since the beginning), so be sure to pay attention to your results before submission if you are concerned about performance.

Happy new US Government fiscal year!

-Greg (@nist.gov)