Cognitec's products incorporate state-of-the-art technologies relevant to working with facial images, in particular deep learning, computer vision, and pattern recognition technologies. Algorithmic optimization since 1995 has resulted in superior independence from facial variances such as pose, mimic, age variance, different hair styles, glasses or temporary lighting changes.
The B12 algorithm achieved superior results in the 2018 NIST FRVT 1:1 with visa photos, and most recently, in the 2019 FRVT: Identification (1:N). The FaceVACS Engine also includes the A16 matching algorithm, featuring a much smaller memory footprint suitable for the development of face recognition applications for mobile devices.
In recent months, our R&D team has been working on an extended suite of algorithms to address today’s face recognition challenges and boost the overall performance of our products.
We are preparing the release of the B13 matching algorithm, including better face and landmark finders, and a new mask detector for alerting to persons wearing or not wearing facial masks. Stand by for product releases!
The algorithm team has also upgraded the age estimator and under-age detectors (18 or 21) for improved accuracy, using a new neural network that works on these three tasks simultaneously.
We are continuing our work on a new matching algorithm for masked faces. Internal tests are showing increasingly confident results. Consequently, we plan to offer a fused engine that combines algorithms for masked and unmasked faces.