Skip to main content

Thales proves top-tier performance in forensic biometrics with AI-improved algorithm

ELFT (Evaluation of Latent Friction Ridge Technology) by NIST is the reference biometric technology assessment protocol related to latent print identification. ELFT evaluates the capacity of a solution to process accurately and efficiently incomplete or low-quality prints – such as those lifted from crime scenes. 

Over the past years, Thales has been steadily investing in its research on latent finger and palm print recognition – and its recent results in ELFT confirms the top-tier performance of its technology. Thales’ latest algorithm is a distillation of innovations and display leading scores on many assessed dimensions. 

With a hit rate of 91.9%* at rank 1, Thales delivers outstanding results. Moreover, Thales algorithm uses less than half the computing power required by other vendors. It is by far the fastest among all participating suppliers’ – comparatively more than 50% faster than any other vendor’s – allowing Thales to deliver excellent accuracy at lower costs and with greater energy efficiency. 


In addition, Thales sees a 44% improvement of its performance on Detection Error Tradeoff (DET), measuring the balance between false positive and false negative. This major enhancement brings sizeable reduction of operating costs – such as unnecessary practitioner time spent manually comparing prints on weak hits.

ELFT results speak to Thales’ commitment not only to provide world-class biometric algorithms – but also innovative solutions with unrivalled performance, built with operational concerns in mind.
Thales is the trusted supplier of dozens of security forces around the world – nearly 100 Thales ABIS are today used for their forensic duties. 

* Cumulative Match Characteristic (CMC) / FBI-Provided Solved Dataset #1 / Probes with EFS Data / Image + EFS. This refers to the percentage of times that the correct subject appears at or above a rank 1 in the candidate list.


biometrics crime scene forensics

Thales Evidence & Investigation Suite