Ritsumeikan Educational Corporation is actively incorporating cyber technology and promoting a new way of looking at school facilities as Campus DX. The school needed to provide an open and comfortable environment while ensuring safety on the university campus, and the challenge was to monitor more than 200 security cameras. In order to build a campus that is more secure and safe than ever before and to study the best security system, Asilla's AI security system will be introduced in April 2023. In this interview, we asked them about the current safety measures at Ritsumeikan University, the background leading to the introduction of the security system "AI Secuity asilla (formerly AI Security System "Asilla")," and the changes and initiatives after the introduction of the system.
●The challenges we were facing
200 security cameras could not monitor all of them and some of them were missed.
●Background of introduction
Seeking to build a new security system using AI in terms of personnel and operation.
●Results after introduction
New awareness through automatic detection, creation of campus safety and security
Company name: Ritsumeikan Educational Corporation
Person interviewed:
・Ritsumeikan Educational Corporation: Mr. Tatsuya Kume, Manager, Contract Section, Finance Department
・Location of the AI Security System: Ritsumeikan University, Osaka Ibaraki Campus
Location where the AI security system was installed: Ritsumeikan University, Osaka Ibaraki Campus
Start of operation: April 2023 (trial operation started in December 2022)
◆interviewer
Group Manager, Customer Success Group, Asilla Corporation
●Many notices by various detections and real-time detection, such as discomfort detection
●Various functions and high detection accuracy
●Easy to use not only for crime prevention but also for disaster prevention
●High commitment to implementation and operation
Kume:
We decided to install AI security cameras at Ritsumeikan University because we felt there was a limit to the number of security cameras that could be viewed by the human eye.
Originally, there were 200 security cameras on campus, but it was very difficult for a human to watch all of them, and even if each camera was watched in one second, it would take 200 seconds.
In other words, if something happened, it would not be detected until 200 seconds later.
Kume:
With 200 security cameras, we were thinking about how to improve operational efficiency. It is inefficient to have a person standing in front of the monitor all the time, and in fact, most of the time nothing happens even if the person is monitoring.
In fact, in most cases, nothing happens even if a person is monitoring the cameras. We were aware of this and were looking for a solution.
Also, even if something did happen, the police would come to us later and say, "Something happened during this time period. Please show me the security camera footage from that time period.
Many of these incidents occurred hours or days after the event, so every time we were faced with such a case, we wondered, "Is this really crime prevention?" I honestly felt "Is this really crime prevention?
Doi:
Yes, that's right. Although there were no violent acts, we frequently found moments of violation, such as the use of skateboards or bicycles in prohibited areas.
However, it was tiring for us to have to deal with such violations, and I think we overlooked some of them.
Kume:
The reason for introducing the AI security system was because I felt that utilizing "AI," with its advanced technology, was the way to go in line with the times.
I thought that the ideal way to use security cameras would be to use AI to automatically send out alerts at the moment of danger and quickly rush to the scene.
I thought that if we could quickly detect danger, we could realize a safer and more secure campus, and that is why I took on this challenge.
Kume:
I had actually researched various products and tried using them on a trial basis, but my honest impression was that the image was different from what I had expected.
In this respect, Asilla immediately brought an actual device and explained it to us while showing us a demo of actual tipping over, etc., so we were able to imagine what it would be like to use the product.
I was able to understand immediately what kind of product it was and how it could be used, so I decided to introduce it first.
Kume:
Kume： Compared to other companies' products, I feel that AI Security asilla is by far the best in terms of detection performance.
Also, we originally had a situation where we could not release the images from the security cameras outside the campus, and some products required us to upload the data to the cloud, but the fact that Asilla's product is server-based and the images are not released outside the campus was also a criterion in choosing the product.
Kume:
The Osaka Ibaraki Campus has no walls and is adjacent to a park, making it an environment where people can enter from anywhere in 360 degrees.
Therefore, the number of cameras is also large, but we decided to start with the Osaka Ibaraki Campus in order to reduce the burden of crime prevention and to strengthen the system.
Kume:
Yes, that's right. Although the campus is open to the general public, we want to make sure that people positively feel that it is an open environment where anyone can come in, rather than feeling uneasy because anyone can come in.
On top of that, in order to strengthen the security aspect, we decided to work hard on expanding security cameras and AI security.
Doi:
It has been about six months since we introduced the system, and asilla has been alerting us to violations and dangers.
I no longer have to switch between multiple monitors all the time, as I used to do before. I like the fact that it alerts me in detail.
Doi:
Of course, if we see a problem at a glance, we will rush to the scene. For example, when we detect a fall, we may see on the screen that the person is really lying down, but he or she may not have actually fallen but is just lying down in that area, or may have fallen down as a joke. If the person is lying in a conspicuous place, we will be careful, but if there seems to be no problem, we will leave it as it is, and people can check what Asilla has alerted them to and make good use of it. It is sometimes a surprise to find out that the system is being used in such a way.
Kume:
I like the fact that AI checks the parts that people used to look at, and by sounding an alert, it changes the situation so that if you only look at that point, there is no problem....
Also, AI is not perfect, so it sometimes makes false detections, but I think that's OK. Even if it's a strikeout, a human can ultimately confirm if it's a strikeout or not.
Kume: That's right. Therefore, I think it is important for AI and people to cooperate with each other, and rather than leaving everything to AI, it is important to first consider how to improve efficiency by incorporating AI into existing technologies. In other technologies as well, I believe that how to coexist with existing technologies, rather than replacing them entirely, will lead to solving issues such as labor shortages. I think it is important to think about how we can integrate technologies that are already in place, rather than replacing them outright.
Doi:
I have been in charge of the operation of "asilla" since its introduction, and compared to when it was first introduced, the number of false positives has decreased considerably. The current status is sufficient, but I expect that the accuracy will improve in various areas. I think it is also attractive that the product is not just installed, but is updated from time to time to improve its accuracy.
Kume:
I am a member of the Finance Department, and in looking through all the contracts, I have noticed that the cost of campus management has been rising year by year. With the ever-changing structure of the campus and the school, I feel that we cannot continue to operate in the same manner.
We thought that if we could reduce the financial burden without reducing the quality of campus management, not only in terms of cost, but also in terms of service to students, we would be able to contribute to the improvement of the school's structure. From this point of view, it was easy to try out ASILLA, which does not require any initial cost.
Kume:
Currently, we have two people monitoring cameras. Currently, we have not made any changes to the number of people monitoring the cameras, but as the number of cameras equipped with the asilla system increases in the future, we expect to reduce resources and change operations.
At the moment, we feel that we are not yet using 100% of the functions of asilla, so we would like to install it in more effective locations and utilize it in a variety of ways other than for crime prevention.
Kume:
In the future, we would like to utilize white cane and wheelchair detection.
If we can ascertain whether such people are present or not, I think it will lead to a greater awareness of the need for careful monitoring.
Also, since the number of people in the building can be confirmed by the headcount function, we think it could be used to respond in the event of an earthquake.
Kume:
We will firmly confirm the use of "asilla" and its effectiveness this year, and by the time the new building is completed next year, we hope to have a full-fledged security system in a form that incorporates AI in a completely new security structure.
ーーーーーーーーーーーーーーーーーーーー
The AI camera market continues to grow at an accelerated pace as vendors compete to improve performance and detection accuracy.
Awareness of AI and robotics is growing, and interest in and understanding of AI security implementation is increasing in a variety of industries.
The accuracy and quality of the system will continue to improve as various companies, including Ritsumeikan University, which we introduced in this paper, share and update their system usage and thinking.
Asilla will continue to cooperate with Ritsumeikan University in its promotion of AI security, and will do its utmost to help the university realize its "Campus DX" vision.
Asilla" is an AI security system based on world-class behavior recognition technology. It converts existing cameras to AI and instantly notifies the user when abnormal or suspicious behavior is detected. This reduces the workload of security guards monitoring video and eliminates the possibility of missing something or overlooking something.
If you are a company or school that would like to solve security issues by DXing, please feel free to contact us.
＜Inquiry URL＞
