evaluation of latent fingerprint technologies (elft)
TRANSCRIPT
Evaluation of Latent Fingerprint Technologies (ELFT)
What are Latents?
Ubiquitous (found on almost all surfaces)
Residual marks from fingers & palms
… but from Very High Value Subjects
Arbitrarily poor quality
Also toe and footprints (all friction ridge skin)
Conventional Fingerprints (not latents)
Rolls
Slaps (Flats)
Rolls
Captured Fingerprints
Flat impression
Latent Fingerprint (powder lift) Rolled Impression
Latent vs. Conventional Prints
Why Are Latents Difficult?
Tips Smudges Weathering
Why Are Latents Important?
Homeland Security
Military & Intelligence Law Enforcement / Criminal Justice
List of Candidates
Output
Rank Subject
ID
Candidate
1 0733091
2 1304033
3 3952340
…
20 0847121
Automated Fingerprint Identification System
(AFIS)
Latent Examiner
Image + features (for comparison)
Ma
rk
Image + features (for search)
Input
Ma
rk
Mark
/ C
om
pare
Latent Examiner
Ma
rk / C
om
pa
re
Mrk
/Co
mp
Match
Latent AFIS Technology
Latent AFIS Technology Problems?
Manual workload …
Indeterminate
Bifurcation
AFIS ID step: Manual Feature Markup
Core
Core
Ridge
ending
Examiner & Workstation
Latent Workstation Screenshot
Latent AFIS Technology Problems?
Manual processing …
Accuracy …
0
10
20
30
40
50
60
70
80
90
100
�Identification Rate (%)
Conventional AFIS
Latent AFIS
37%
Conventional vs. Latent ID Rates
Latent AFIS Technology Problems?
Manual processing …
Accuracy …
Interoperability …
AFIS Interoperability: Now
AFIS Interoperability: Future
Core
Indeterminate
Incipient
Pore Core
Ridge
ending
Protrusion
Bifurcation
Improved Feature Quality o ridge quality map
Improved Feature Set: o endings/bifurcations
Dot
o pores
o protrusions
o incipient ridges
o dots
o creases
o scars
o skeleton
Extended Feature Set (EFS)
What is ELFT?
• Open evaluation of Automated Latent Fingerprint Identification Systems (AFIS) using automatic feature extraction and matching and standardized features hand-marked by human experts (EFS).
• Interactive effort between NIST and latent AFIS community to improve accuracy, promote interoperability, and reduce reliance on human examiners.
Acquire Latent
Matchers (SDKs)
Configure Hardware
Compile Latent Test Sets
ELFT Testing Architecture
o Execute 1-to-Many searches
• Image-only searches
• Examiner-assisted searches
(image + feature markup)
• Operational images
• Extended Feature Sets
o Measure & Analyze Results
• Accuracy
• Selectivity
• Resource requirements
• Gap analysis
Iterate process
1. Evaluation Reports
2. Feedback to Standardization
3. Technological Gap Analysis
4. Reference Data
Evaluation Protocol
2006 NIST Latent Fingerprint Testing Workshop
2007 ELFT Phase I Evaluation
2008 ELFT Phase II Evaluation
2009 NIST Latent Fingerprint Testing Workshop ELFT Phase II Miss Analysis Sessions ELFT-EFS Public Challenge
2010 ELFT-EFS Evaluation #1 ELFT-EFS Miss Analysis Sessions
2011 ELFT-EFS Evaluation #2
ELFT History
Measured Accuracy Improvement
(ELFT-EFS 1 vs. 2)
Customers and Community
Industry Academia
Developers & Researchers:
• Measurement & failure analysis
• Reference data
Government Law Enforcement
• Standards
Users & Practitioners:
• Performance rankings
• System integration & tuning
• Standards
• Conferences, Journals
• Conferences, Workshops
Path Forward
Develop best practices to reduce manual workload • lights-out processing of some operational workload • searching and encoding strategies • quality metrics
Evaluate state-of-the-art latent fingerprint and (new) palmprint AFIS • interoperability testing (LITS/EFS profiles) • matcher and data fusion • candidate reduction techniques
Expand scale of ELFT testing
•operational scale tests •operational vs. research algorithms
Questions?