However, it’s important for security leaders to keep in mind that all video analytics are not the same. Unlike conventional video analytics that employ image recognition technology to recognize static objects and classify objects, advanced behavioral analytics employ AI-driven technology to autonomously detect, analyze and recognize actions and events.

Next-generation video analytics is video understanding

Behavioral analytics recognize the behavior of diverse objects and their relevant actions and context. For example, if a vehicle is driving down a road and suddenly stops, this may indicate a problem if the vehicle stops in the middle of the road, yet may not be a problem if the vehicle stops on the side. Behavioral analytics automatically identify the location of the action as being in the middle of the road or the side of the road.

Other common examples of behavior analysis include the ability to differentiate between individuals embracing or fighting, recognizing a calm peaceful crowd or a raucous crowd which may become violent, and identifying a person falling versus a person bending over to tie their shoelaces. The ability to distinguish a potentially threatening or dangerous event of interest from a normal occurrence provides security professionals with the unique ability to initiate fast and appropriate responses to deescalate or minimize damaging incidents or even prevent such incidents from occurring.

Behaviors and dynamic situations have many variations that need to be understood and recognized. People don’t fight in only one way — they fight in many ways, and people can fall on the floor in many ways. In the examples above, behavioral analytics must discern between events that look similar, such as fighting or hugging and a person falling to the ground or simply bending over. Behavioral analytics can do this by learning from video clips to create a unique signature for the behavior in the clip. Then, when processing live video streams, it can spot such signatures and raise an alarm that something is amiss.

Security leaders can leverage analytics to proactively address safety concerns. Behavioral analytics can have the ability to autonomously detect events of interest, such as violent and suspicious activity, crowd behavior, perimeter (protection) violations, public health violations, traffic congestion and accidents, environmental and personal safety threats, and occupancy and mobility counting. Video analytics can improve security and environmental, health and safety monitoring accuracy by eliminating the need for personnel to monitor large numbers of cameras simultaneously. As a result, they can also help reduce personnel cost associated with monitoring large video systems and responding to false alarms while improving security operations center (SOC)/security personnel performance by allowing them to focus on real events of interest and provide additional services.

Automating intelligence with surveillance analytics

Behavioral recognition video analytics provide actionable intelligence on the events that matter most to help minimize potentially dangerous situations, reduce liabilities, mitigate risks, and maintain compliance with new and emerging mandates.

Behavioral recognition video analytics are already being deployed around the world for a wide range of use cases including safe and smart cities, transportation hubs, banking and financial institutions, corporate and education campuses, healthcare facilities as well as for industrial and manufacturing occupational and environmental safety. The success of established and future deployments will only serve to further accelerate the deployment of advanced behavioral recognition video analytics, transforming traditional reactive video systems into proactive sources of video data and intelligence.