January 18, 2023

Developing new sensing technologies to contribute to the evolution of AI

Researcher, Biometrics Research Laboratories
J. Kenji Clark

Born and raised in Canada, Dr. Clark moved to Japan, his grandmother’s homeland, for his doctorate. After completing a doctorate program in September 2019, he joined NEC in April 2020 after working for half a year as a postdoctoral researcher. His current research in optical coherence tomography allows him to make use of the knowledge of nanophotonics and optical materials that he gained during his university research.

Recognizing the fingerprint "behind” the fingerprint

I currently research a technology called optical coherence tomography (OCT) at NEC's Biometrics Research Laboratories. The technology involves capturing 3D images of the sub-surface of objects using the coherence of light. Originally, it was a technique that was used to examine the structure of the retina in the field of ophthalmology. Our team is working on applying this technique to fingerprint recognition. If we can develop fingerprint recognition using OCT, then we would be able to scan not only the fingerprint on the surface of the skin, but also scan the dermal fingerprint pattern beneath the surface fingerprint. Using the dermal fingerprint pattern enables high-accuracy fingerprint recognition even when the skin surface has scratches or wrinkles, not to mention that the detection of fake fingerprints would also be possible.
Fingerprint recognition algorithms have been one of NEC’s strong points—the company has been ranked number one in the world* several times in fingerprint recognition benchmark tests hosted by the U.S. National Institute of Standards and Technology (NIST). Once the sensing device that we are currently researching and developing comes to life, we can keep enhancing the accuracy of NEC’s fingerprint recognition technology and further expand its application in society.
Our research results are periodically organized into research papers for publication. I have presented two papers at academic conferences during the three years of being with NEC. Currently, I have one paper going through peer review. Looking at the researchers around me, it seems that most submit one or two papers to academic conferences a year at NEC. While being a corporate research center, we see active participation in academic conferences here at the Biometrics Research Laboratories.

Free and progressive environment unlike the common image of a “Japanese company”

After completing my doctorate, I aspired to invent something that can make a difference in people’s lives, so I started job hunting. While I visited several companies, NEC’s recruiters all gave me the impression that they were very kind. Instead of speaking to me from the standpoint of “just trying to recruit good employees,” they discussed with me topics such as what research I am interested in and were concerned with me finding a laboratory that matched my interests. That attitude was so pleasing that it became one of the main reasons I decided to join NEC.
When I actually started working for NEC, I was surprised by how the company recommended a very flexible working style. When my friends abroad found out that I got employed by a Japanese company, they warned me that relationships with bosses would be difficult. That never happened here, and I am working in a forward-thinking style of work environment. Taking advantage of the company’s Super Flex System, I can come to and leave work flexibly. A lot of colleagues are remote working. Relationships with supervisors are also equal and neutral, allowing for unreserved opinion exchange.
Above all, the laboratory hosts many researchers with various backgrounds and experiences. I can learn approaches in other fields and vice versa—attempt solving issues in other fields using my techniques. Through such interdisciplinary interchanges, I can expand my personal knowledge and accelerate innovation.

Advancing AI with sensing technologies

When I do research, I am constantly mindful of expanding my own knowledge. Once I find an issue, I work to not only find and apply solutions, but also clearly understand the essence of that issue. After I find a solution, I try to understand the essence of the solution as well. Then I can discover new issues and think about my next research theme. Being preoccupied with simply applying solutions to issues you have currently can cause uncertainty in what to do next. Developing a deep understanding allows you to go further—after solving the issue, new ideas come up for better accuracy and possible applications to other areas.
For example, my research involves improving the accuracy of OCT images using a signal processing technique called compressed sensing, which was quite distant from my specialty and therefore difficult to understand at first. While I could have improved the OCT image accuracy by just utilizing the technique, I took a longer way around and studied for long hours about compressed sensing by reading numerous research papers. Consequently, I came up with a better way to use compressed sensing in OCT, which led to the next step in my research.
My current goal is to develop a practical dermal fingerprint pattern scanning device. At the end of the day, my essential goal is to provide new data for the advancement of AI technologies. For example, while image recognition uses camera images for analyses and creates various values, typical camera images have limitations on what can be analyzed. In order to obtain more information than what camera images can give, we need to develop new sensor technologies. As a researcher of sensing technologies, I aspire to contribute to the development of more powerful AIs through developing innovative sensors that provide access to new information.

A day at work

Message to my past self in my school days