IoT Boosts Video-based Security Systems

Security is a top priority for companies and government institutions across the globe, and that is fueling demand for artificial intelligence-based video surveillance systems. Strategically placed in smart cities, airports, and federal buildings these systems can ensure that citizens, employees, and passengers are safe.

Image credit: Atos

The global market for video surveillance systems is projected to reach $81.3 billion by 2028, with a CAGR of 9.5 percent. IP video surveillance systems, which lead the market, was valued at $27.3 billion in 2020. The commercial market is the largest segment, valued at 14.68 billion, and is growing amid increased security concerns.

Companies like Atos are taking advantage of this market opportunity. Paris-based Atos, a leader in cybersecurity, cloud, and high-performance computing in Europe, acquired Ipsotek in June 2021. Ipsotek developed the VISuite AI-based video analytics software platform, which enables Atos to integrate a broader array of security measures across its solutions.

Insights from Deep Learning Object Attributes

AI video analytics can be edge or server based, says Anthony Valle, senior presales engineer at Ipsotek. Edge computing happens at the camera level. It focuses on clear-cut items, such as objects in the frame, changes in the scene, or heat maps. Those solutions are sufficient in certain applications. However, camera-based detection solutions are limited by what they can identify and processor power.

Ipsotek’s VISuite is a server-based AI video analytics solution that uses powerful, deep learning GPUs. VISuite not only detects objects in the camera frame, but it also can identify attributes related to those objects, providing greater insight to security operators.

VISuite AI centers on its patented Scenario-Based Rule Engine (SBRE), which enables the software to follow objects and behaviors. It uses video analytics, artificial intelligence, facial recognition, distributed sensors, and other multimodal information to differentiate between typical and risky behaviors under surveillance. VISuite can trigger alarms based on those behaviors.

Ipsotek’s VISuite AI works using a large network of cameras, which track objects in real-time as they move from camera to camera. Advanced trackers and AI detectors collect metadata on each object in the frame, including the object’s size, shape, appearance, speed, and trajectory on a geomap. That and other data feeds into the SBRE, which automatically assesses the risk level and triggers follow-up actions.

Ipsotek's VISuite platform

Image credit: Ipsotek

Easy Deployment and Customization

To simplify large-scale rollouts, VISuite AI has automatic scene calibration capabilities. The system collects data for a brief time and then establishes parameters needed for calibration and calculates the GPS coordinates of objects based on those of the cameras. Ipsotek’s library of tested rule templates simplify configuration. Rule templates can be directly applied to cameras directly, either individually or to groups of cameras through virtual camera zones. Each camera’s field of view can be adjusted to focus on detection zones.

VISuite AI is an open platform that can be integrated with third-party cameras, video management systems (VMS), and physical security information management systems (PSIM). API support enables integration with additional partners.

Atos, a strong player in the European market, has partnered with Arrow Electronics to expand its US footprint. “We view Arrow as an extension of our efforts in the Americas. Arrow is a tech partner that brings more value to what we can do for the end customer,” says Emilio Taddio, vice president of business development for cybersecurity and BDS for Atos North America.

“Together with Atos on the technology side, we are developing a market in the US for a solution that has been around [internationally] for a long time,” says Roland Ducote, director of supplier programs in Arrow’s Intelligent Solutions division.