For a surveillance camera to be truly effective in times past, it required at least two elements: a system to manage multiple cameras at once, and a human being watching that system to respond to threats and questionable activities. Since then, many practical and useful enhancements have come to market, so much so that even the smallest businesses can now protect their investments, employees and property to the fullest, without investing human capital on hours spent.
Very recently, Google has pioneered the use of deep convolutional neural networks to identify objects in static images with a very high degree of accuracy. They achieved this by using a multitude of finely networked and highly specific processing units tied to massive data storage containers, combined with dense replicating algorithms. What this means in layman’s terms is that each computer taught itself how to identify photos by holding on to every bit of data about that photo, sharing it with other computers, and then comparing the differing results. You may already be using a reduced version of this software, if you are using facial recognition to log into your mobile phone or a specific app.
Current IP-based camera video is ultimately nothing more than a lot of individual photos that are sequenced together to create a video that’s really just a moving set of pictures. It’s only been a matter of time that this same kind of machine learning could be used to identify specific features of a human face, or even the shape of raised letters on a vehicle’s license plate, in real time monitoring of a video stream and then compare those results to a database of known quantities – a very specific example being a former employee who may be barred from entering a restricted area gets captured by a camera and his face is compared to a record of employee ID badges for the past 5 years. And while an organization is still investing in a physical security presence, using an automated facial recognition software system frees up a large amount of man hours, effectively becoming a force multiplier and allowing less staff to achieve more.
Another way that advanced video analytics provide value outside of security is for visitor metrics, put simply how many different individuals are entering your space at any given time. This can be incredibly valuable for retail stores, transportation infrastructure, or large public spaces such as sports stadiums or convention centers to measure human traffic throughout a space. A well rounded video management system is rapidly becoming a must-have for any organization, and as facial recognition and video analytical software become better, they also become more cost-efficient as well. If you’ve ever wondered about how to get the most out of your camera network, or to have a new camera network protect your facility, contact us today.