Traditional face recognition has allowed for security watch-list recognition at terminals and critical infrastructure, white-list recognition to detect intruders, and recognition across intelligent camera networks.
The Suspect Technologies team has built a highly advanced, unconstrained facial recognition application for these use cases. This involves finding and recognizing faces in various lighting conditions, rapid motion and shakiness, and various face sizes and colors.
Real Time Alerts
Facial Recognition WatchList Alert
The system can be used both historically and in near real-time. The code is infinitely parallelizable, meaning a 20 hr video could be processed in under one hour even. Detection on average takes about 100ms per image and is 99.19 % accurate on leading benchmarking databases. It works on almost any system environment and with any video stream
The technology makes it possible to enroll subjects automatically through video capture, and works correctly even with partial facial concealment, wearing of glasses, handkerchiefs or caps, changes in facial expression, difficult lighting conditions and slight rotations of the face.
Suspect FR is optimized for difficult environments in movement, such as airports, metro and train stations, shopping centers, sports stadiums and urban centers.
Real time, with a searching speed of > 1,000,000 faces per second
High accuracy with glasses and most sunglasses
Not affected by a mustache, beard hair, make-up, hats, etc.
Support of different permissions according to accesses and restricted areas
Modular and scalable software
Automatic subject enrollment function
Allows for multiple watch-lists (blacklists / white-lists)
Runtime alarm management, highly tuneable Compact, flexible system with multilingual management
Alarms can be exported to PDF format and visualized on remote mobile devices