IntelliVision Technologies Corp says that it has created “another level of control and management” for security systems integrators with the latest version of its facial recognition technology. The new version now runs as a cloud-based service as well as on edge devices and on-premise servers.
“Face recognition is becoming a must-have security feature for smart homes and buildings, smart retail, transportation, access control and the gaming industry,” said Krishna Khadloya, Vice President of Engineering and Products at the US-based manufacturer. “Our face recognition now gives systems integrators the flexibility to implement facial detection and recognition in the cloud, on servers, or at the edge, providing real-time results and reducing bandwidth usage.”
IntelliVision is part of Nortek Security & Control and has a focus on AI and Deep Learning video analytics software for Smart Cameras, providing video analytics solutions for markets including Smart Home/IoT, Smart Security, Smart Retail, Smart Business, Smart City and Smart Auto (ADAS). Its analytics products are in use in over 4 million cameras around the world and are used to provide actionable insights to home, business and security systems.
Facial recognition is part of the IntelliVision family of video analytics solutions that also includes object classification (humans, vehicles, pets, aeroplanes), intelligent motion detection, intrusion/perimeter watch, object left/removed, license plate recognition (ALPR/ANPR) and audio/sound recognition. The latest version of IntelliVision’s facial recognition technology is optimised for in-camera processing and is reportedly capable of detecting at least 10 faces in less than 1s with only 25% CPU usage on an Ambarella S5L-based camera. Up to 20 faces can be recognised in a single frame. In the cloud-based Face Recognition as a Service (FRaaS), recognition time is sub-second and allows for additional analytics in the overall workflow where camera processing capacity is limited.