Deep learning and artificial intelligence (AI) have helped the security monitoring industry transform video clips into actionable intelligence. These technologies enable effective transformation of live or recorded video into structured metadata that can deliver quantifiable insights. However, our research indicates that many customers have concern over AI vendors’ claim on significant false alert reduction.
Computer vision AI has the ability to monitor large amount of video feeds in real-time, but one issue that has plagued many users is these video clips are analyzed separately as they come through. With advancement in deep learning, camera AIs are now so advanced that they can differentiate between situations that trigger different actions.
AIs with situational awareness can now attain a comprehensive spatial-temporal understanding of a site’s environment. Products like Sentry AI perceive images and metadata surrounding a clip in its corresponding time and space, rather than analyzing it as an isolated view. AIs can now organize, digest, and interpret situations based on clip history, past alarms, reference images, etc. Mimicking neurological processing, situation-specific AI algorithms and thresholds lead to unparalleled response accuracy and help security organizations become more cost-effective and agile.
We talked a lot about how AIs can help security organizations, but we want to put data into your perspective and say why this is important for you. From a report by SDM on security integration revenue by service, we see a steady decline on revenue from installation. Instead, monitoring services have been growing at a fast rate. Historically, AI’s benefits might have not been emphasized as much in one-time revenue generated from installations. However, as the market moves towards long-term monitoring, AI’s accuracy, cost effectiveness, and ROI increase will become a differentiating factor.
North American security systems integration sales revenue 2014-2020, by service (Source: SDM)
Written by Adrian Xin