Brad McMullen on AI Breakthroughs in Video Monitoring

SDM spoke with Brad McMullen, president, security products and solutions (3XLOGIC, PACOM, Sonitrol) at Securitas Technology, Uniontown, Ohio, about the effect that AI is having on monitoring from the manufacturer’s perspective. Ahead, McMullen answers a few questions about how AI has shaped the design of 3xLOGIC’s solutions.

SDM: How are you designing your systems/solutions to support monitoring centers using AI?

McMullen: At 3xLOGIC, our VIGIL VMS is designed to support monitoring centers by shifting the burden of AI processing to the camera edge. 3xLOGIC Edge-Based Deep Learning cameras are equipped with object-based person and vehicle detection to accurately identify and classify events, so they only send targeted notifications to monitoring stations. This edge support eliminates the need for monitoring centers to deploy or manage their own AI systems which, in turn, reduces complexity, cost, and false alarms. Monitoring center operators will only receive relevant actionable events, so response times can be shortened.

SDM: What specific AI features have proven most effective in real-world monitoring operations?

McMullen: The most effective AI feature has been edge-based deep learning classification of objects — especially the ability to distinguish between people and vehicles, but also through filtering out irrelevant motion activity. This dramatically reduces false alarms and ensures that monitoring center operators focus only on true events that require their attention.

SDM: How important is localized AI (adjusting alerts by time of day, geography, type of site) to your product roadmap?

McMullen: The ability to set schedules, object or vehicle classes, sensitivity or confidence of detections allows users to tailor their protection to their needs, filtering out normal activities that don’t require a response to effectively capture those events that are legitimate anomalies.”

SDM: What do you believe will be the biggest AI-enabled breakthrough in video monitoring in the near-future?

McMullen: I expect natural language search of video footage to be the biggest AI-enabled breakthrough in video monitoring in the near future. Being able to ask a system to “show me when a person loitered in the parking lot for more than 20 seconds after 3a.m.” will make search faster, easier and more efficient.”