Belmont FRP
Belmont FRP (Facial Recognition Platform) performs automatic face detection, recognition, and further search operations utilizing advanced, modern algorithms.
​
The system highlights faces, creates neural samples, and compares these samples with faces stored in the database. A person’s access is then automatically managed, or operators are notified in case of an alarm.
Unlimited database
​FRP has no limits for database records or detection and recognition processes.
Detection and recognition angle
Recognition is guaranteed at an angle of 35° to the horizontal and vertical axes. However, in practice, FRP functions even at angles of 90°.
Recognition in challenging environments
Various facial accessories, expressions, and styles such as masks, hats, glasses, smiles, closed eyes, raised eyebrows, makeup, beards, or moustaches can all pose a problem for facial recognition processes. FRP overcomes these challenges and reliably conducts recognition.
Low-quality images
​Video cameras used for facial recognition do not usually provide the ideal quality, especially in challenging lighting conditions. Regardless of this, the recognition engine must work with the images available. FRP successfully operates even with quality degradation resulting in facial features lacking visibility and being poorly presented.
Anti-Spoofing
The internal Anti-Spoofing algorithm determines the difference between a live face and a face printed on paper or displayed on an electronic device, thereby preventing false recognition.
Rapid performance
​The system can rapidly search a database of 1,000,000 events in less than 3 seconds.
API
We provide REST API and continuously support it with up-to-date releases of the latest FRP software packages. API may be used to acquire and integrate data from FRP to 3rd party systems.
VMS and Camera agnostic
FRP does not depend on any VMS or any specific brand of camera. The only requirement is an RTSP stream to provide images of a scene, so that recognition can be conducted.
Age changes
The human face inevitably undergoes physical changes due to aging. Cosmetic surgery can create artificial alterations in physical appearance. This could result in a discrepancy between stored images and an individual’s face in real-time images. FRP maintains its reliable recognition process and resists drastic facial alterations.
Tracking
The tracking feature for recognized individuals is available across cameras situated in separate zones on the site or on multiple sites. An individual’s location and path between locations can be tracked, giving the time and location of each instance of recognition.