RealNetworks
The SAFR AI platform for facial recognition is designed to work economically with high performance and fast processing, to instantly detect and combine millions of faces in near real time, even in difficult conditions where faces move, at different angles, in low light or partially darkened.
The system offers an accuracy of 99.87% and offers results 3-5 times faster than other facial recognition algorithms, capable of detecting a face moving through live video and recognizing it in less than 100 milliseconds. It also has a feature to improve accuracy, which allows you to group the different forms of appearance of a person, including face masks, adding them manually.
Currently, with the widespread use of face masks, the system is "learning" to improve its occlusion logic, allowing health professionals to walk through safe areas without removing their protective elements, but retaining their ability to recognize security threats.
SAFR can be deployed on a single computer to monitor multiple IP cameras or scale to thousands of cameras in a distributed, locally hosted or cloud architecture.
Request FREE information about this product. Please include the URL of the desired product.
Leave your comment