Multi-Camera Facial Recognition Security System
The Face Recognition Company to demonstrate multi-viewpoint machine-learning system at International Security Expo
The Face Recognition Company (FRC) announce its multi-camera face-recognition security product at the International Security Expo in London.
The company, which has been in stealth mode for three years, has developed a very accurate facial recognition security system, optimised for commercial businesses. Available in January 2020, its first product is designed to provide enhanced security for retail and commercial premises or similar environments. The system can be installed at entrances to identify and alert about shoplifters on a watchlist, for example.
Incorporating multiple cameras, machine-learning algorithms and a user-friendly mobile app for providing alerts and managing watchlists, the system is designed for ease-of-use by anyone from store management to security staff. In real-time, it detects unwelcome visitors by recognising faces that have been flagged previously, before alerting security or management via the app.
Current facial recognition systems typically access CCTV feeds, and are frequently complex, bespoke installations focussed on national security, policing and big transportation infrastructure.
“Today’s facial recognition algorithms are very accurate but can only be effective when they have an image of suitable quality and perspective,” said Tim Noest, CEO, The Face Recognition Company. “Grainy CCTV images or faces captured beyond a certain angle from the front view reduce the accuracy significantly. With our multi-view system, we are focused exclusively on capturing high-quality facial data, greatly increasing the chances of making a positive match.”
The FRC system comprises three components:
- Cameras. FRC has developed a patent-pending multi-camera technology. A typical setup incorporates five high-resolution cameras trained on one capture zone – a doorway entry into a retail store, for example.
- Mobile app. Users control and interact with the system via a simple app. Managers with appropriate security approvals have full system control and can set up and manage watchlists. Store staff and security can receive and respond to alerts but have no access to watchlists.
- Back-end server. The local server runs the image data processing and machine learning in the background.
Users don’t interact with the server beyond the initial setup phase.
The FRC system is not designed to be a replacement for existing security infrastructure and processes such as CCTV guards. It is designed to augment security operations to run more effectively and efficiently.