THE WEEKLY DETAIL... the internet newsletter for Latent Print Examiners.

UPDATES ON CLPEX.com

Updated the Fingerprint Interest Group (FIG) page with FIG #73; Distortion: Doubling, by Sandy Siegel of Texas. You can send your example (anonymously if you desire) of unique distortion through Charlie Parker: Charles.Parker@ci.austin.tx.us. For discussion, visit the CLPEX.com forum FIG thread.

Updated the Detail Archives
_________________________________________

we took a look at new Quick Capture fingerprint technology from the FBI CJIS Division.

_________________________________________
NITS ELFT-EFS - Evaluation of Latent Fingerprint Technologies: Extended Feature Sets

Purpose
The purpose of this evaluation is to determine the effectiveness of human latent examiner marked fingerprint features on latent fingerprint search accuracy, specifically with respect to the comparative accuracy of image-only searches, image+minutiae searches, and image+extended feature searches.

Background
Recent work in latent fingerprint matching has focused both on increased automation (image-only latent matching), and extended features:

• The NIST ELFT evaluations and operational AFIS have shown that image-only latent searches are practical, with some promising fully automated image-only latent matching results — but an evaluation needs to be conducted, using new data, that investigates the comparative effect on accuracy of image-only and feature-based searches, and can provide data to determine what characteristics are likely to result in accurate automated searches.

• A (draft) standard method of defining a comprehensive set of friction ridge features has been proposed through the work of CDEFFS 1. This Extended Feature Set (EFS) specification allows examiners to define what they see in an image — but an evaluation needs to be conducted that investigates the comparative effect on accuracy of EFS searches over traditional minutiae searches. It is critical to note that these features have a cost in examiner time: human latent expertise is a limited resource. Latent fingerprint examiners have marked minutiae for AFIS searches for years. Asking latent examiners to spend to the time to mark up a new feature for an AFIS search is only worthwhile if there is a clear benefit with respect to matcher accuracy.

This evaluation seeks to address the needs of decision makers to determine how to optimize efficiency and accuracy when conducting image-only, image+minutiae, and image+extended features latent searches.

Path Forward
Summary of path forward:

A. Public challenge dataset (SD27-1000 – see below) is already available, without Extended Feature Set (EFS) markup.

B. Public challenge dataset with juried EFS feature markup will be made available. The dataset will be released incrementally as markup is completed.

C. Comments will be solicited from potential participants with respect to the overall impact of EFS on matcher accuracy, comments on the potential usefulness of the various features, and comments on methods of marking the EFS in the public challenge dataset.

D. A workshop will be held to provide input to NIST in the planning of a challenge evaluation of matcher algorithms using various combinations of EFS features. If appropriate, the EFS specification and/or the guidelines for its use will be revised in response. We are tentatively planning this workshop to be held in February 2009.

E. A challenge problem will be conducted, at the participants’ facilities, using the public challenge data, with self-reported results.

F. A workshop will be held to discuss the results of the challenge problem, lessons learned, and to provide input to NIST in the planning of the actual evaluation. If appropriate, the EFS specification and/or the guidelines for its use will be revised in response. If appropriate, the public challenge dataset will be revised in response and made public.

G. Based on the findings of the challenge problem and workshops, NIST will conduct an evaluation of latent feature matching, using participants’ software on NIST hardware at NIST facilities. The evaluation dataset will be sequestered data, marked up in accordance with the findings of the challenge problem. The images in the evaluation dataset will be different than the ELFT Phase II dataset.

Public challenge dataset (SD27-1000)
A sample dataset of latent and exemplar images, with juried human markup of most or all EFS features will be made available before the test. The public (challenge) dataset will be based on the SD27 dataset, which has been rescanned at 1000ppi (now described as SD27-1000). The SD27-1000 latents with juried minutiae markup (not other EFS features yet) were included in the ULW 5.4 release (August 2008). The full 14-image 1000ppi exemplar sets for SD27-1000 are currently available – but without feature markup. These can be sent to interested parties within the next few weeks if requested.

Juried feature markup entails multiple expert human latent examiners conferring to create an “ideal” markup that is as objective as possible.
For the challenge problem, the gallery (background) data will include the rolled 10-print sets from SD27 (mates of the latents, at 500 or 1000ppi), along with SD29/SD30 as unmated background, for a total of approximately 450 subjects. Exemplars for the gallery will be images only. Different tests will be run with 500ppi and 1000ppi exemplars. Note that SD29 and SD30 contain the same images from 216 subjects, only differing in that SD29 was scanned at 500ppi and SD30 was scanned at 1000ppi.

Proposed outline of challenge problem
The following is a proposed outline of the planned challenge problem:

• The challenge problem will involve 1:N searches using latent 1000ppi images provided with human groundtruthed (juried) markup of CDEFFS features.

• Exemplars for the gallery will be images only. Different tests will be run with 500ppi and 1000ppi exemplars.

• The challenge problem will be conducted at the participants’ facilities, using the public challenge data, with self-reported results.

• Different tests will be run for the following search types:
o Image only
o Image with minutiae
o Image with minimal/various sets of CDEFFS features (comments requested on what this would involve)
o Image with complete set of CDEFFS features
o Some tests may evaluate performance of features without image.

For further information
As additional information becomes available, updates will be made at the ELFT-EFS website (fingerprint.nist.gov/latent/ELFT-EFS), with notifications sent to interested parties.

To register as an interested party, please send contact information (name, affiliation, email, telephone) to latent-efs@nist.gov.

1. ANSI/NIST Committee to Define an Extended Fingerprint Feature Set — see http://fingerprint.nist.gov/standard/cdeffs

_________________________________________

Feel free to pass The Detail along to other examiners. This is a free newsletter FOR latent print examiners, BY latent print examiners. There are no copyrights on The Detail content and as always, the website is open for all to visit!

If you have not yet signed up to receive the Weekly Detail in YOUR e-mail inbox, go ahead and join the list now so you don't miss out! (To join this free e-mail newsletter, enter your name and e-mail address on the following page: http://www.clpex.com/Subscribe.htm You will be sent a Confirmation e-mail... just click on the link in that e-mail, or paste it into an Internet Explorer address bar, and you are signed up!) If you have problems receiving the Detail from a work e-mail address, there have been past issues with department e-mail filters considering the Detail as potential unsolicited e-mail. Try subscribing from a home e-mail address or contact your IT department to allow e-mails from Topica. Members may unsubscribe at any time. If you have difficulties with the sign-up process or have been inadvertently removed from the list, e-mail me personally at kaseywertheim@aol.com and I will try to work things out.

Until next Monday morning, don't work too hard or too little.

Have a GREAT week!