Neural networks
We develop innovative facial recognition algorithms
based on deep-learning neural
networks.
They operate many times faster
than traditional algorithms.
Comparison and recognition
against a database of 50 million images
per second on a single processor.
Solutions for any facial recognition task
Biometric non-cooperative face in the crowd recognition
Instantaneous facial recognition in a video stream with high reliability.
Compatible
with all IP cameras.
Recognition based on social media photos
The lowest requirements for image quality, adaptation of algorithms to low
quality
images. Processing of databases with an unlimited number of
images and search for
matches in open sources and social networks.
Facial expressions, make-up
do not have a significant impact
Acceptable
angle
of the face
Minimum
distance
between pupils
Effectiveness and accessibility
High-speed calculations based on GPU and CUDA technology have significantly reduced hardware
platform
requirements for the facial recognition system.
Reduction in cost and rapid development of the accuracy of algorithms
has made facial recognition accessible for a wide range of users.
Facial recognition systems based on our technologies are used in the
retail and service sectors, sports venues, banks, transport hubs, logistics
enterprises and other areas.
Furthermore, recognition systems can differ greatly from each other
depending on the tasks and conditions of operation.
It can be a small
stand-alone device, a distributed system with hundreds of connected cameras and a
dedicated data center,
a software suite integrated with other applications, or a
cloud service with an API.
Facial recognition device:
the size of a credit card