Asilla, Inc. is aiming for a world where incidents and accidents are prevented using behavioral recognition AI as its core technology. Here, we would like to write about a world in which the behavior recognition AI complements and reinforces what is lacking in security cameras, thereby providing another dimension of crime prevention value.
Security cameras are on the increase. They have become so familiar that you can often see signs saying that security cameras have been installed when you walk down the street. Discussions have even begun on making security cameras in train cars mandatory (*1).
The increase in the number of security cameras seems to have a certain deterrent effect. Even so, people continue to intentionally or accidentally cause incidents and accidents, and drastic solutions have yet to be found.
The list is endless.
If there are so many security cameras, it would be possible to prevent incidents and accidents by checking their images 24 hours a day and responding immediately, but this is not realistic, as you can't just stand at the monitor all day long checking images.
However, this is not realistic, and it is not feasible. That sounds great. I would love to have such a system.
So as a trial, we tried an existing security camera (with motion detection) in our company. This camera is designed to send notifications to your smartphone when it detects any movement, which seems convenient.
But in fact, it is not.
The camera's built-in motion detection function detects when something moves and notifies you, but this is not what I was expecting,
・When a person gets up from his/her seat or sits down.
・When the lights go on or off in the room.
・The calendar flickers in the breeze from the air conditioner.
For example, it notifies me of things that I don't need to be notified of, such as people getting up and down from their seats, lights turning on and off in the room, and the calendar flickering in the wind from the air conditioner, which is extremely annoying.
In the end, I immediately changed the notifications to OFF because there were so many notifications (but that doesn't make sense).
As this example shows, what people really want is a smart security camera system that notifies them "only" when "something unusual or strange" occurs.
We wanted such a smart security camera system. Fortunately, we have AI technology that has been refined over the years (our AI technology is capable of analyzing people's movements in chronological detail at a practical level). By optimizing this technology in the context of crime prevention security, we have perfected it into an AI security system for facilities called "Asilla" and released it in January.
Asilla, an AI security system for facilities
・Immediate notification only when abnormal or uncomfortable behavior occurs.
・Autonomous learning of normal behavior in a given situation to detect unusual behavior.
・Processes a large number of camera images with a small amount of computational resources.
We have already received positive feedback from customers who have installed this product, saying "This is the kind of product we have been looking for," and the number of companies interested in this product is increasing one after another. We hope that you will try our product and experience its greatness. If you have any questions, please feel free to contact us.
The products we have released this time still have great potential beyond that, and we plan to continuously evolve them in the future. Based on your reactions, we foresee a future in which a new world of security camera systems utilizing behavior recognition AI will unfold. We hope you will continue to look forward to our company's efforts to realize a world in which camera images are successfully utilized to prevent incidents and accidents before they occur.
Source:
