Having people continuously watch surveillance camera video to detect anomaly is not just inefficient but can also lead to anomaly being overlooked. Therefore, to realize high-quality and high-efficiency service, SECOM has built a system in which a computer automatically detects anomaly, while humans, using their high ability to make judgments, check the results.

We have long been working on image recognition technology to detect anomaly from surveillance camera images. In 1998, we used this technology to develop SECOM AX, a remote imaging on-line security system. In 2010, we developed the Intelligent Emergency Alert System, which automatically detects and reports robberies using surveillance camera images. In 2014, we developed the Walkthrough Face Recognition System, which enables face recognition to be performed on persons walking through the shot. In particular, pattern recognition*1, which forms the core of artificial intelligence (AI), is an important component technology. Since long before the advent of deep learning*2, we have provided society with services that incorporate AI technology.

In December 2017, at the international passenger terminal building of Haneda Airport, we participated in a demonstration experiment sponsored by the Ministry of Land, Infrastructure, Transport and Tourism for an advanced security system aimed at preventing international terrorism. We have contributed to strengthening security and reducing the burden on security guards in crowded environments with technology that automatically detects suspicious behavior from surveillance camera images via image recognition, and then ascertains characteristics such as clothing and body shape to track individuals, as well as technology that detects suspicious items such as packages abandoned for a certain time.

In recent years, as the need has grown for wide-area surveillance at large-scale events, we have also been conducting research on “crowd analysis technology,” which analyzes the behavior of entire crowds. With this research, we conducted a demonstration experiment at a marathon in Japan and contributed to reducing the burden on security observers by estimating crowd congestion and analyzing crowd movements.

*1　Pattern recognition: The process of selecting and extracting objects that satisfy certain rules or have certain meanings from huge amounts of information, including images and sounds

*2　Deep learning: A method whereby a computer learns from large amounts of data, in order to automatically identify rules and features

Sensors, which stand at the forefront of our services, are indispensable for tracking the status of people, objects, spaces, etc. To improve our services, we need to develop sensing technologies to allow us to assess situations more rapidly, more reliably, and more precisely.

To that end, we are conducting research via two approaches: advancing sensing technologies individually and advancing these technologies by integrating multiple sensors. In terms of individual sensing technologies, we are continuously developing technologies related to light, electromagnetic waves, audible sound, and ultrasonic waves. In terms of integrating multiple sensors, we also aim to establish a technology that compensates for the weaknesses of each sensor and maintains high performance in various environments.

One possible application of this research is the monitoring of homes. To achieve this, we face the major challenge of realizing technology that accurately ascertains situations, while predicting future anomaly from slight changes and also protecting privacy. For this reason, we are researching technology that uses sensors for ultrasonic waves as well as sensors for electromagnetic waves, such as microwaves and millimeter waves, in order to track the detailed movements of people and objects without using cameras.

For more than 50 years, SECOM has been realizing the potential of the concept of the Internet of Things (IoT), which aggregates sensor signals through communication lines. To bring this concept to further fruition, we have researched the security of the communication infrastructure that transmits and receives signals.

All kinds of things are connected via the Internet, and while its convenience is increasing, a wide variety of risks in cyberspace exists, as well as new threats such as cyber attacks, which are continuing to evolve daily. To secure safety on the Internet, it is essential that we ensure the confidentiality of communications to guarantee the reliability of systems and services, maintain security to assure the authenticity of data, and protect user privacy. Accordingly, SECOM is working on encryption, signatures, and authentication as core security technologies, while also conducting research in cyber security. Furthermore, to realize advanced and personalized services utilizing Big Data, we are researching high-speed, high-reliability network infrastructure technology and working on distributed processing technology.

Precisely analyzed security plans are essential at events such as international conferences and sports competitions, where VIPS, top athletes and spectators from around the world gather. We have been researching on a technology for utilizing spatial information that integrates geographic information system (GIS) and building information modeling (BIM) technologies. In 2016, we have launched this technology under the brand SECOM 3D Security Planning.

By precisely reproducing on-site environments in digital spaces and assessing on-site situations to identify and analyze risks, we can drastically reduce the effort necessary for on-site surveys. Furthermore, utilizing 3D maps and simulations would enable us to check terrain, buildings, and other facilities intuitively. Also, it is possible to study omnidirectional images taken from moving vehicles, in order to efficiently assess geographical environments and make various plans.

To envision and create services needed by society in the future, we are researching analysis and optimization technologies for security planning and operations so that quality, efficiency and eventually the service itself can be improved. Concurrently, we are researching simulation technologies for processes which are difficult to physically formulate, such as human and service processes.

In our simulation technology research, we are investigating methods to model and predict human movements, aiming to optimize operational plans while providing safety and comfort, even for crowded areas. In addition, through the use of service operation simulations, we are investigating methods to provide insight into staff headcounts, work shifts and service quality.

To address the challenges associated with a super-aging society, we are conducting research on data analysis technology in hospitals to improve the quality of medical care and management efficiency, as well as research on preventative care and health promotion.

In 1994, SECOM began offering Hospi-net, Japan’s first remote image diagnosis support service. In particular, while CT and MRI are rapidly becoming more commonplace, the number of specialists capable of conducting image diagnosis remains small. This has led to a social issue, in that diagnoses require lengthy times, resulting in delays in treatment. In response, the SECOM Group’s medical, technology development, and information communication specialists were assembled to research and develop remote image diagnosis based on technical infrastructure such as communication technology and image compression technology cultivated in SECOM’s on-line security systems.

In addition, we used communication robots to conduct demonstration experiments with the aims of maintaining and improving the quality of life (QOL) of elderly people at SECOM Lifestyle Partner Kugayama in Suginami City, Tokyo. As a result, we confirmed that helping elderly people take medication and alleviating feelings of loneliness have certain effects. For elderly people to live with peace of mind, it is important that they not only receive hospital medical services and regular nursing care services but also live healthy and fulfilling lives at home. We are conducting technical research for health services that match users’ lifestyles, such as guidance for taking medication that corresponds to their lifestyle rhythms.