What is the ELFT algorithm evaluation by NIST?
The Evaluation of Latent Friction Ridge Technology (ELFT) by NIST is a well-recognized name in an important benchmarking by a third-party entity to evaluate the performance of latent fingerprint identification algorithms.
Latent fingerprints are either partial, distorted, or low-quality impressions that were inadvertently deposited on a surface. Usually found at crime scenes or on pieces of evidence, latent prints provide valuable evidence to link individuals to suspected criminal activity. Algorithms must accurately and efficiently evaluate these challenging impressions in order to provide reliable and transparent forensic results.
How ELFT Works
ELFT by NIST is designed to establish an objective, standardized means to measure the accuracy and effectiveness of latent friction ridge impression examination algorithms. These impressions emerge from highly textured skin and may appear as prints from fingers and palms. When used to test algorithms, the application focuses on the extraction of features, creation of templates, and comparisons against reference databases. By providing common test datasets, ELFT allows for independent testing of different technologies. Evaluations are imperative as part of the procurement process.
How NIST Measures Algorithm Performance and Reliability
NIST publishes analysis reports that detail not only matching accuracy but also computational performance. Matching accuracy includes false positive and true positive identification rates. These rates directly measure the probability of associating an unknown print with the right identity. Executing speed and hardware scalability, as well as software flexibility, are examples of metrics used to measure… Resource investment… prospecting for the ferruginous testing methodology, etc. This allows end users to understand and define the trade-off between identification speed and reliability.
Evaluations are performed with controlled datasets and predefined procedures. This allows for the reliability and complexity of each problem that the end user is trying to address. The results of the evaluation process are evaluated by NIST, are consistent, repeatable, and can be used for future repeating purposes. However, NIST does not rank manufacturers or vendors as winners or losers, so that data can be used as a reliable tool. The metrics obtained by stakeholders allow them to formulate new product and deployment plans in a neutral way.
Key Benefit of Third-Party Examination by NIST: Transparency
Transparency sets ELFT apart. NIST publishes all test results with an individual summary report. Independent Search & Harness examined all the vendor reports here. Forensic laboratories, decision-makers, and policymakers can access and review test results. By providing information, NIST understands the critical process of achieving a healthy, burgeoning industry.
Leveraging Trust and Compliance with NIST ELFT Assessments
Vendors whose latent friction ridge impression examination algorithms were evaluated can use the independent results in product documentation. Independent NIST testing can be used to support vendor product stochasticity, allowing customers to validate a claim through an independent source. Stakeholders ready to learn more here about integrating NIST examination reports, which can be used to support trust, can find more information in the Force Factor information here.
All in all, ELFT is a great way to maintain an evidence-based healthy skepticism about fingerprints, with a protocol available to quantifiably reduce that skepticism. It’s a win-win.
