Discover How Realtime Behavioral Recognition Analytics Detect and Predict Violent Activity

January 25, 2021 2:50 pm

people fighting

Video surveillance is an established and proven tool implemented across all industries and organizations for its ability to help create and maintain safe and secure environments. In recent years, intelligent video analytics technology has increased the functionality of surveillance beyond passive monitoring with the ability to recognize specific types of activity to help mitigate threats. For example, conventional vehicle detection analytics can identify vehicle color and its type, or if there is movement in a facility where there should not be any activity.

However, conventional analytics do not have the ability to understand what is happening in any given situation. For example, two people who are physically close together, moving from foot to foot and intermittently coming into contact with one another could be dancing, embracing, or involved in mortal combat. Conventional analytics only detect their presence without any interpretation of intent.

There are many cases like this for which the human brain applies context to understand what is being observed. Early video analytics did not have the capability to perform this function. However, through Artificial Intelligence (AI), machine learning and deep learning, the most advanced video analytics can detect and understand many behaviors that represent security threats. This new discipline, behavioral analytics for video surveillance, is helping organizations around the world prevent and mitigate violence, theft and other risk.

Here are three key behaviors that today’s best video analytics can recognize. If your solution doesn’t include these, you should be looking for something better.


1. Video analytics to identify people fighting

As mentioned, the action of two people fighting could appear to analytics to be similar to other actions such as dancing or hugging. Advanced behavioral analytics can discern between these actions and actual fighting, defined as at least two people who are hitting each other. The software identifies hitting as hands moving towards the other person, touching and then disengaging, with the action being repeated numerous times. Such a system can also recognize differing fighting styles including kicking, punching or wrestling in both indoor and outdoor settings.

The action of a person stabbing another person can be detected by advanced behavioral analytics as well. In this case, there would be two people standing close together, with one of them performing a stabbing movement. It is not necessary for the knife or other object used to stab to be visible.

2. Video analytics to detect weapons

Advanced AI can be trained to recognize a person holding a weapon in a threatening position. The presence of the weapon alone, not being held by a person in a threating position, would not trigger an alert. The behavioral analytics detect the combination of the position and the existence of the specific weapon.

3. Video analytics to predict crowd violence and contextual loitering

Large groups of people can behave in a wide variety of ways, most of them non-violent. A highly advanced behavioral recognition system can learn to recognize suspicious behavior that can turn into violent action in a crowd. For example, AI-based video analytics to predict violent behavior may trigger an alarm when a large number of people in a group begin to push one another or otherwise act aggressively for a specified period of time., enabling action before the situation erupts into a riot.

This software recognizes both objects and context. People will move and behave differently at a bus stop than they do near an ATM. The combination of human behavior and location type delivers insights to classify the movement either as loitering (by context) or as something different. For example, a person who may be loitering near an ATM would be flagged as suspicious activity triggering security personnel to watch the situation or inquire if assistance is required. This combination of behavioral analysis sets a new benchmark in video analytics as a preventive versus a monitoring solution.


AI-based analytics enable automated understanding of video content

The new advent of intelligent behavioral analytics dramatically elevates the capability of video surveillance to reduce threats and risk. By identifying events of interest based on behavior, and autonomously alerting security personnel with actionable insights, it is possible to leverage the vast and growing pool of surveillance video to create a safer environment.


Learn more on how viisights innovative behavioral analytics can transform virtually any video monitoring system into an intelligent security and business intelligence solution.

Go Back

Blog Posts