Video analytics can provide a much-needed boost to security in schools, turning reactive devices into proactive systems.
December 14 marked the tenth anniversary of the Sandy Hook School shooting, one of the deadliest active shooter incidents in US history. Twenty first-graders and six educators were killed on that day.
The incident prompted calls for better security at schools. But a decade later, school shooting incidents continue to remain a major concern for the US. In May, an attack on an elementary school in Uvalde, Texas killed 19 children and two teachers.
The physical security industry has offered several solutions that schools can make use of. However, given the situation, it is more important to first understand why schools are unable to stop gun violence and then help them overcome the challenges.
Why school shooting incidents persist?
According to Dean Nicolls, CMO of Oosto there are three root causes behind the continued school shootings. Beyond the broader issues of mental health and gun control, the first issue is ensuring that weapons never enter school grounds.
“This can be accomplished through a combination of physical security measures – metal detectors, video cameras, door locks, visitor management systems, and intrusion alarms,” Nicolls said. “Unfortunately, many schools are strapped financially and cannot afford some of these basic physical security safeguards.”
The second obstacle is having an effective communication strategy once an active shooter is on the premise. But one of the primary sources of failure for all school security systems lies not in the technology itself but in the element of human delay.
“According to Bruce Montgomery from Honeywell Security, ‘It’s critical that solutions providers find ways to remove the delay in reporting violent incidents to law enforcement and faster systems that omit single-point failures’,” Nicolls added. “The third stumbling block is to ensure that there is a person (or a small group of people) who have the requisite authority to act. Schools need to know who are their "go-to" people – people trained on how to respond to active shooter situations – and can respond quickly without further escalating the situation.”
Mike Lahiff, CEO and co-founder of ZeroEyes pointed out that school shooting incidents remain a problem because most school security measures are reactive, not proactive. If law enforcement is not called until after the first shots are fired, it is too late.
“Furthermore, school shootings tend to result in what we call the 'fog of war,' with multiple people calling 911 with contradictory information, so that first responders have no idea what type of scenario they are actually walking into,” Lahiff said. “As a result, precious time is lost. The average active shooting lasts about 6 minutes but given all the conflicting information from 911 calls, first responder response times have taken as long as 18 minutes.”
The root cause is a lack of actionable intelligence. For the most part, current security cameras are not being leveraged to proactively identify a threat before shots are fired - they are used mainly for forensic evidence, which is after the event.
How video analytics can help schools
There are a few types of video analytics that can potentially help counteractive shooter scenarios. Weapons detection is a type of AI-based video analytics software that “watches” surveillance camera feeds and sounds an alarm when the machine vision thinks it sees a person holding a gun. The challenge is that guns are usually concealed so these types of analytics have limited applicability.
“A second type of video analytics, real-time facial recognition, watches these same video feeds from CCTV cameras and looks for people perceived as potential threats and placed on a predefined watchlist,” Nicolls said. “The watchlist can include ex-students, students, or former employees who are prone to violence or even registered pedophiles. If any of these individuals set foot on school property, security guards can be alerted. However, this, too, has limitations; if the active shooter is a current student in good standing, then they would not be on a watchlist, and no alarms would sound.”
A third type of detection is gunshot detection. Though this is technically not video analytics, the technology relies on a system of sensors to detect the sound of a gunshot, transmitters to send a message to the police dispatch center, and a computer to receive and display that message.
“Video analytics can be used to detect guns the very moment they appear on a security camera so that trained personnel can alert law enforcement and on-site security to the threat before shots are fired - possibly even before the shooter enters the building,” Lahiff said. “Analytics can also help provide first responders with information about the shooter's appearance/clothing, location, and weapon type.”
How integrators and end users can leverage analytics
It’s clear that school security is thorny and complex and that there’s no one right answer. But it’s also clear that hardening a school’s security needs to be at least one prong of the solution. “Unifying things like cameras, access control, doors, HVAC, and fire systems” is critically important, according to Guy Grace, K-12 Unified Security Consultant at Apple to Apple. “Things like facial recognition enhance your ability to detect, deter, and deny” bad actors from entering school grounds.
Nicolls pointed out that while people argue for stronger gun controls and better mental health programs, integrators can work with schools to design real-world solutions that exploit the power of today’s technologies. He listed four key areas that systems integrators can work on.
Social Media Monitoring: The FBI and local police can monitor public social media posts in real-time to identify threats and respond quickly. These services cannot detect private one-to-one text messages, but they remain an important tool that should be leveraged.
Gunshot Detection: Gunshot detection systems use acoustic sensing technology to identify, discriminate, and report gunshots to the police within seconds of the shot being fired. A gunshot detection system usually involves a number of sensors to detect the sound of a gunshot, and transmitters to send a message to the police dispatch center.
Weapons Detection: Weapon detection systems use high-performance object detection algorithms using deep neural networks. These AI algorithms work with your existing video surveillance systems to automatically detect knives, guns, and firearms using deep-learning models.
Watchlist Alerting: Many schools in the US have lists of known security threats. This often includes ex-students, parents, and faculty prone to violence and sexual predators who live in the immediate vicinity.
Simultaneously, integrators can help schools think through processes that determine the location of the active shooter and how to effectively and appropriately communicate the threat to teachers, students, law enforcement, and parents.
Final word
School shooting incidents continue to remain a major problem, especially in the US. Despite several initiatives, not much has been done to put an end to them. Traditional security solutions help to some extent, but clearly, they are not enough.
Video analytics can provide a much-needed boost to security in schools, turning reactive devices into proactive systems. With the right budgetary allocation and measured steps, AI-enabled solutions can make a difference in this space.