The results from the fall of 2020 U.S. Department of Homeland Security (DHS) Science and Technology Directorate’s Biometric Technology Rally are being reported and for the second year in a row, Rank One Computing (ROC) is reporting a strong performance.
The DHS Rally is used each year as a way of assessing the capabilities of biometric algorithms in the identification of passengers in real-time screening systems. This year, due to the COVID-19 pandemic, the ability to identify individuals wearing masks was added to the evaluation.
The ROC algorithm was the only one among all those tested to achieve an accuracy of greater than 90 percent on masked faces when using two different acquisition systems, and had the overall second highest score with an accuracy of 91.9 percent.
The news of the strong performance comes during what has already been a busy start to the year for ROC. In the first week of January the company received a US patent for its proprietary passive biometric liveness detection technology that uses micro-texture analysis to confirm the authentication of an image.
A few weeks after the new patent announcement, ROC also announced the release of a new facial recognition plugin for the Milestone XProtect Video Management System (VMS).
“In addition to the accuracies achieved by the ROC algorithm, it continues to be a stalwart of efficiency in the NIST FRVT benchmarks,” notes a blog post on the company’s website. “The combination of top-tier accuracy and hardware efficiency both put Rank One in a category of its own and save ROC’s partners and customers substantial amounts of money in hardware costs.”
ROC’s facial recognition algorithm also outperformed several iris recognition algorithms that took part in the Rally, out-pacing them in accuracy with an average score of 98.1 percent, and by a margin of 91.9 percent vs 67.3 percent when identifying individuals wearing masks.
March 19, 2021 – by Tony Bitzionis