Innovatrics ABIS gets high marks on NIST ELFT benchmark

Latent fingerprints are the ones forensics experts collect and analyze, those carefully captured traces of human mischief that criminals leave behind at crime scenes. A newly published benchmark of companies that provide algorithms for identifying them ranks the biometrics firm Innovatrics among its top performers, according to a company release.

The EU-based independent provider of multimodal biometrics tools fared well on the NIST ELFT benchmark. Its Innovatrics ABIS contains algorithms for fingerprint biometrics, as well as face and iris.

Many people know about latent fingerprints from TV police procedural dramas, and while details tend to be grossly exaggerated in fiction, real-world detectives who hunt for the elusive prints still face significant challenges. Latent fingerprints from crime scenes, says Matus Kapusta, Head of the ABIS unit at Innovatrics, “are usually partial, captured in uncontrolled environment, on different backgrounds and with diverse capturing techniques.”

The Innovatrics ABIS, says the release, can make identification faster and easier, in part because of how it uses AI to generate partial synthetic fingerprints based on the latent fingerprint dataset. However, the company is careful to mention the limits of synthetic fingerprints.

“One has to be extremely cautious when using them in testing and product validation,” says Jan Lunter, the CEO, CTO and Founder of Innovatrics. Lunter has previously published comments on the same theme, arguing for the benefits of synthetic data in biometric algorithm training for the purposes of cutting bias and making data more sharable, but cautioning organizations not to abandon data drawn from real life.

“Synthetic data can only mimic reality,” he wrote in a recent article printed in Biometric Technology Today. “Therefore, real-world values may be missing.”

Article Topics

accuracy  |  biometric testing  |  ELFT  |  fingerprint biometrics  |  Innovatrics  |  NIST  |  synthetic data