Deploying CV to detect Mass Shooters Early On

Photo by Sandra Grünewald  on Unsplash


The blogs have been written by the Revca team with the help of a countless interns that have also contributed to bringing these points to you.

Follow us on:

In twelve years, 1368 people were shot to death in the United States in mass shootings. Over the past couple of years, the number of people who have died has increased despite institutional safeguards to manage mass shootings. Associated with this is the physical and mental trauma of the people affected by this. There are many facets to this crisis and multiple views on how this should be tackled. Here, we attempt to provide a solution that uses artificial intelligence to proactively combat gun violence. 

The success of controlling a mass shooting event is dependent on many aspects, of which the most important is the response of the authorities. Damage control depends on how fast authorities are notified and how long it takes for them to reach the site of an incident and enforce an action plan. This currently is largely done manually, by affected individuals notifying someone of the danger, who may be the concerned authority or may in turn get law enforcement involved. 

This chain of operation is, however, very much prone to error. In the case of an active shooter scenario, people may find it difficult, or even unable to send out a distress signal. At times, attempting to notify anyone during an active shooter may even put you in danger. This is the communication gap that we attempt to solve by introducing computer vision. 

Why Do We Need to Change

Some time ago, CCTV cameras were a novelty. However, over the years, such cameras have become ubiquitous in our surveillance and security systems. However, while introducing CCTV was a large step forward, it is time we take the next step towards security. But this requires answering two questions. “What is wrong with current systems?”, & “How does computer vision help solve this problem?”. 

Current security systems are surveillance based. They only act as extra eyes, and as a platform for recording. Monitoring such footage requires one, or multiple individuals working round the clock. This has multiple issues. In the case of large areas, it is practically unable to constantly monitor all footage at all times. The chance of human negligence in such a work is extremely high. This is the first scope of error in current surveillance systems.

Secondly, these surveillance systems cannot act proactively. In the case of a live shooting, or any unfortunate incident, it cannot notify anyone on its own. This requires an individual, as mentioned earlier to act on the data that is being collected. However in the case of a shooting or so, if the responsible individual misses the incident due to his negligence, or is neutralized, such systems cannot help at all.  

What is Computer Vision

Before we understand how computer vision, we need to understand what is Computer Vision.

Computer Vision is a form of Artificial Intelligence that derives information from the input visual. In simple words, It can look at an image of an apple, and understand that the image is that of an Apple. The first instance of Computer Vision was developed at Dartmouth University, where a summer intern worked with AI to develop software to read handwriting. 

Computer Vision works depending on how we teach it. It learns to differentiate an Apple from an Orange by shifting through humongous quantities of images of Apples before it is ready to accurately identify an Apple.  Currently, though, Computer Vision is expected to surpass $48.6 billion by 2022.

In the present day though, most of us use multiple iterations of Computer Vision in our daily life. This ranges from Face Locks, image-based shopping, face detection in file management software, and so on. It is the same technology, in different iterations that we shall deploy to enhance security, in this case, to deal with mass shootings.  

Why Integrate Computer Vision?

Now that we have a ballpark idea of how Computer Vision functions, we shall understand how Computer Vision shall overcome the above-said shortcomings. 

Integrating Computer Vision into surveillance systems makes your surveillance; in short words; intelligent. Computer Vision monitors that data in real-time, looking for patterns, or objects that have been registered as harmful; or unfavourable. In the case of mass shootings, the most unfavourable object is the Weapon. Computer Vision software can hence be programmed to detect unholstered weapons. Detection is the primary activity, where the aim is to overcome human negligence in monitoring.

We have in theory already overcome the second, which was the inability of surveillance systems to work on their own. Now the question, is what if the persona being notified has already been neutralized? This can be overcome with multiple people being notified, some of whom may function outside the institution. To illustrate this, imagine a mass shooting in an institution. Normally the person monitoring surveillance would be functioning from inside the institution. However, if the whole institution is compromised, there is little this individual can do. Hence having contact points physically away from the institution is crucial in notifying concerned authorities. 

One major point of concern in this is the chance of false positives. This again depends on how we train the model. The chance of false positives reduces depending on the inputs we train the model upon. There have been models for gun detection, which are trained effectively so that there have been Zero false positives to date. 

An extra step could be added to the model to prevent false positives. The model could be designed in a way that it first notifies a certain individual in the case detects unfavourable activity. In case the first notification is not attended to within a certain time frame, multiple people including the Remote Point of Contact are notified. Such a system may be considered to deal with false positives, in the uneventful case that false positives occur. 

Bottom Line

In shorter words, Computer Vision provides you with a system of surveillance that operates with a level of autonomy, which can be controlled provided you are authorized for the same. This means you get a surveillance system that assists the human monitoring while acting in the unfortunate absence of a human person to direct it. 

At Apture, we believe that our command over technology should be utilized for the public good. We believe that children should be able to go to school without the trauma of their classmates being gunned down and that people should be able to go to public spaces without the fear of being killed or wounded by bullets.  If you are concerned about the safety of your students, or if you are part of the civil society that believes in safer neighbourhoods, contact us to know more about how you integrated our solutions with your surveillance systems. 

Get Subscribed to our Newsletter to stay on top of recent CV developments!

Related Articles

The only platform you need to implement CV without the hassle.

A Context based CV deployment platform. Build Intelligence into your Cameras with Machine Learning. Derive insights from your visual data to drive growth.

Contact Info

Subscribe Now

Don’t miss our future updates! Get Subscribed Today!