ALUMNI ALUMNI

The joy of pursuing a new research topic and discovering a new world

During my time at the Laboratory of Media Dynamics, I focused on the pass, which is an important element of soccer strategy, and was researching a method for automatically determining effective pass courses based on player locations that were obtained from video. This allows computers to understand elements of the game, for example, the team that is controlling the game as well as the key players. Through my research, I wanted to create an environment that analyzed the strategy of soccer videos and then displayed information such as who’s controlling the game, as well as the game’s development to the user, enabling them to greater enjoy watching the match.
Because there were extremely few existing works in this area, I discussed many new hypotheses with my supervisor, and verified them. Despite experiencing many setbacks, we managed to create the field of pass course analysis, and were happy to that even people in the industry were interested in our approach.
Each student in the laboratory had a wide variety of research topics. I think this improved the quality of the laboratory’s research. Through seminars, we were able to discover new techniques from different fields, and discussions enabled us to expand our field of vision. Furthermore, everyone was actively working towards the same goal by submitting to domestic and international conferences and going to various exhibitions. I’m very grateful for the opportunity to spend my university years in such an environment.

What I learned at the laboratory is the foundation of who I am today

Right now, I’m working for KDDI FRANCE, and am living in France. The image of KDDI in Japan is predominantly mobile and fixed telecommunications, but outside of Japan, KDDI focuses on providing integrated IT solutions to customers through servers and PC.
It was during my time at the lab that I first became aware of the possibility of working outside of Japan. Through reading papers written by authors around the world and submitting my own work to international conferences, I started to think of working overseas as something that is natural. When working overseas, there are many differences to Japan – for example, the business environment, relationships with customers and partners. It is important to understanding these differences, as well as to provide services and conduct negotiations in a manner that is suitable to the particular country. Sometimes it can be difficult, because there are many things one needs to learn, but working with a different culture or environment can also be very fascinating.
I’ve often used know-how from the laboratory in the course of my work. For example, I had to analyze the market or internal company environment for a customer, construct a hypothesis and then think of an optimal solution. This is quite similar to the “analysis, hypothesis, proposal” process that we learned in the laboratory. This way of thinking that I learned at the laboratory has been extremely useful. The same thing applies when pushing a project towards a deadline – it’s fairly natural if you have experience in conducting research in order to present it at a conference.
When working overseas, I feel proud of Japan’s technology and its high quality. On the other hand, I also feel that the products used by the global society correspond to only a small proportion of our potential ability. Onward, I want to go beyond the framework of a telecommunications company, cooperate with a wide variety of industries, bring Japan’s high-quality products and services to the world, and contribute to creating a convenient and pleasant society for the world’s inhabitants.

Full of constant discussion, a lively environment that facilitates growth and development

During my time at the Laboratory of Media Dynamics, my research involved recognition of videos captured by surveillance cameras, in particular, pedestrian detection. This research allows one to develop a security system that automatically detects suspicious individuals from surveillance camera videos, and sounds an alarm. The research consists of video capture, feature extraction via image processing, pedestrian classification via pattern recognition. By working in this direction, not only did I manage to pursue my research interests, but also obtained fundamental knowledge in the relevant technical areas.
I remember the laboratory as being very lively, with an atmosphere rich in discussion regardless of seniority level. This environment allowed me to delve deeply into my own research, and brush up on my research skills through discussions with other lab members. On the other hand, the research directions given during seminars and paper reviews were extremely strict, and there were many opportunities to present at conferences and publish to journals, so my ability to write clearly improved significantly. Writing ability is of great value no matter where you work, so I think it became a very important asset for me.
Furthermore, the laboratory actively pursues conference presentations and projects outside the university. In my case, I went to international conferences in Okinawa and Las Vegas, and was able to obtain invaluable knowledge through discussions with people outside the university. Finally, on top of their own research, students of the laboratory are responsible for administration of lab infrastructure such as computers and networks, as well as supporting Web systems for conferences. The skills I gained in performing such administration tasks remained extremely useful even after I left the laboratory.

Convert research results into products, contribute to a safe and secure society

Currently, I’m involved in the research and development of biometrics. Biometrics utilizes physical characteristics such as fingerprint or veins to establish an individual’s identity. Recently, biometrics have found applications in a wide range of fields. Examples include vein sensors in ATMs, and fingerprint sensors in smartphones. Hitachi is currently focusing on “finger vein authentication” and holds a significant share of ATM and industrial systems.
Biometrics brings together image sensing, feature extraction via image processing, and individual classification via pattern recognition. The required technical elements are exactly the same as what I used for pedestrian detection back at the laboratory. In my current work, I’m involved in research that is closely related to that of the laboratory, and converting that to products makes me feel very worthwhile.
Due to the spread ICT in our society, many systems around us are becoming more convenient. However, the risk of new forms of crime and terrorism is also increasing. Since nobody wants to use a system that is prone to risks, it is necessary to utilize technology that improves security, such as biometrics.
Furthermore, biometrics don’t just improve security, they also improve convenience, and allow new services to be created. Biometrics can be used for purchasing goods, opening locks, submitting administrative paperwork – opening a whole world of services you can use completely empty-handed. I find this idea very exciting.
Therefore, my current goal is to spread technology that I developed, and contribute to the creation of a more secure and convenient society.

Knowing “how to do research properly” remains a valuable asset even today

While I was at the Laboratory of Media Dynamics, I was researching a fractal block encoding method based on genetic algorithms. “Fractal image compression” reduces the compressed image size by utilizing the self-similarity properties of the image. My research utilized genetic algorithms to quickly find parts of the image that are similar.
I joined my current company because I saw a dream in which the image processing technology that I learned at the Laboratory of Media Dynamics traveled through the underwater fibre-optic cables and was being used all over the world. Shortly after I joined KDD, it became KDDI and shifted its main focus to mobile communications, and the possibility to realize my dream grew even more. Furthermore, the ability to not only sow the seeds of fundamental research, but to engage in a wide range of fields such as technology development, planning, and product development is extremely fascinating.
Currently, I’m working on research and development of technology for supporting future communication carrier business. In my case, I’ve worked on many projects over the last 5 years. Most recently, I’ve worked on big data analysis, sensor data mining and augmented reality.
Furthermore, I’ve seen my research results used in business planning and strategies. Recently, we’ve made a series of TV commercials with the theme of “Turning Surprise into Common Sense”. Some of my own research results were used in these commercials. My specific work is different to what I used to research as a student, but knowing “how to do research properly” is still a common theme. I learned the fundamentals of research as a university student, and it remains an invaluable asset to this day.

I’m glad to have chosen this laboratory. I thought so when I graduated, and I still think so now.

For a long time now, the Laboratory of Media Dynamics has place extremely strong emphasis on inter-student connection, both horizontally and vertically. When I joined the laboratory as a third-year undergraduate student, there was an established tradition for all lab members to go and eat lunch together. The lab environment enabled lively communication between students, and you could receive advice from senior students working on other topics, even outside of seminars. No other lab had such traditions. I clearly remember thinking “I’m glad to have chosen this lab.” This continued even after I graduated, and I still meet with past graduates around twice a year, in Tokyo, for drinks with fresh graduates. The members of the Laboratory of Media Dynamics aren’t just friendly – they’re also working in some very fascinating industries. One day, I’d like to form a team of LMD members and work on a cross-industry project.
Furthermore, I think that in the future, the sciences (research, technology development) and the humanities (marketing, business and sales) ought to be closer together. I want to create an organization that can deploy research or business processes by integrating the sciences and the arts, and optimizing everything across the board. I believe that is connected to the development of communication. In order to achieve this, I have a present goal to supplement my Engineering Ph. D. with an MBA, and recently, I’ve started attending a management graduate school.

We’ve shown off our research achievements at trade fairs. I can feel the progress.

My research topic at the Laboratory of Media Dynamics was related to the clustering of acoustic signals.
Recently, due to the spread of hard disk recorders and blue ray discs, people have begun using record a large volume of TV programs, and then viewing only the interesting parts, for example, viewing only the scenes that their favorite artist is singing in. In order to improve this functionality, it is necessary to automatically analyze the video signal, divide it into selectable scenes, and index them, in advance.
The previous example has two requirements: an image index that corresponds to the artist, and a sound index that corresponds to singing. My research focused on the latter – automatically creating an index that can discriminate between different parts of the audio signal such as singing and talking. There was another student who was researching music similarity, and by combining our research results, we created “Easy Finder” – a recommendation system for similar music and videos. We demonstrated this system at the largest domestic trade show – CEATEC Japan. This was part of the “Information Grand Voyage” project by the Ministry of Economy, Trade and Industry. It gave us an invaluable opportunity to show our research achievements to professionals in the industry and people in general. I felt that it really expanded my world.

You can apply your laboratory experience to projects in the real world

Currently, I’m in charge of development of a recognition algorithm at a car manufacturing company. Recently, cars have become equipped with driver assistance systems that detect potential collisions with other cars or pedestrians. My job is to develop an algorithm that utilizes internal car sensors, more specifically, cameras.
Compared to my student days, the volume of data I work with now is magnitudes greater. Cars are on the road regardless of weather conditions: during the day, at night, rain, snow or shine. Worldwide. Naturally, research and development requires large video volumes. Subject to various conditions, we install cameras into vehicles, get them to drive thousands of kilometers around the world, and then obtain the recorded data. When you look at all this data, sometimes you think “How am I ever going to classify all this?” Nevertheless, the way to approach the problem is exactly the same as back in graduate school. Look closely at the data, make a hypothesis, confirm it via an experiment, confirm the result, and then move on to the next problem. Rinse, dry, repeat.
Furthermore, when I was at the Laboratory of Media Dynamics, I would often gather my daily research results into a technical paper and present them at domestic and international conferences. Once I started working, I could see the large difference that this experience made when I compared myself to my colleagues. For example, when writing business report, in order to communicate the contents to your superior or colleagues, you need to organize the content and results support its validity with sufficient data, and create an interesting presentation that will captivate your audience. This is exactly the same process as gathering research results into a technical paper. I’m most confident that my experience at the Laboratory of Media Dynamics has had a profound effect on my way of working.
Presently, manufacturers across the industry, not just automobile manufacturing, are speeding up their development processes. If you focus only on the Japanese market, you will be left behind. Perhaps you can say something similar about academic research. When looking out into the world, I want to expand my abilities to their maximum potential, and to create things that can contribute to people around the world.