Setting themes in accordance with individual interests
Acquiring the basics and learning the approaches of predecessors and
identifying analytical methods that contribute to clinical medicine

Health Sciences
画像情報学分野 教授
Medical Image Computation Professor
増谷 佳孝
Yoshitaka Masutani

増谷 佳孝

Improving the accuracy of ML-based MRI analysis from a methodological perspective



●Please tell us the concept for your research.

My research mainly involves MR imaging and its analysis, especially obtaining highly precise quantitative values expressing tissue characteristics based on various models and machine learning (ML), with the aim of contributing to clinical practice. For example, with diffusion MRI it is possible to predict changes caused by diseases and tissue microstructure from the behavior of water molecules, but large quantities of imaging data is required and the prediction accuracy is also affected by noise included in the images. We use ML to resolve this problem. In concrete terms, our research centers on methodology for resolving problems in learning by synthesizing large quantities of learning data using models comprising MRI signals and imaging simulations incorporating added noise (Fig. 1). These research results are made publicly available as freeware (refer to end-of-sentence-related link: Fig. 2). In addition, we also conduct research on various systems related to processing and analyzing medical images (Fig. 3).



●What is the atmosphere like in your lab?

The lab was only established six months ago, and we are focusing on building the foundation necessary for beginning research. Students are starting with essays and programming practice. Essays are written on themes comprising their own interests or “favs”, topics with which they are extremely familiar, in order to provide training in written expression in preparation for writing technical papers. Programming is learnt by way of “make10” algorithm implementation examples for finding formula for producing 10 using four arithmetic operations when given four numbers between 0 and 9 (e.g. 1, 1, 5, 8 ⇒ 8/(1-1/5) = 10). We also allocate research themes based on students’ interests and academic path.

プロセスなくして結果なし まずは飛び込んでみることが重要
No results without process: What is important is to just jump in!



●What is your message for students who wish to join your lab?

The more you learn about the subject, the more questions you will have. You will also gain much inspiration through learning about the different thoughts and approaches of your predecessors. As a result, you will devise your own problem analysis methods and approaches, and conduct research through trial and error. This is enjoyable, but certain processes are necessary for arriving at that point. Results cannot be achieved without processes. In other words, it is important to just jump right in your field of interest. There days it is very easy to obtain information, including from the internet, so I encourage potential students to utilize this information and research this research field as far as they can to make sure that it could be a field in which they are interested. If you have any questions, I am happy to answer them. Accordingly, we welcome anyone who has even the slightest interest in the field of research conducted by our lab. Come and do research with us!

Career path possibilities for students completing the program and expectations for their contributions



●What kind of career paths are there for students who complete your program? In what fields are graduates active?

As of now, this lab has only been operating for six months, so we have no graduates just yet. However, I expect that graduates who obtain national certification, such as radiology technicians, will play active roles in clinical practice while maintaining a research mindset. Graduates will probably also carry out research on questions or issues they encounter in their day-to-day work and present papers at academic conferences. Depending on their performance, graduates could also have careers working mainly in R&D at medical device manufacturers or research institutions.



●It is possible to study at your lab while working at a hospital or business enterprise?

In my previous position, I supervised adult graduate students who worked as radiology technicians or at manufacturing companies, and I also had the experience of conducting joint research with radiology technicians enrolled in the graduate program. Depending on the student’s motivation, I think that it is entirely possible to study at this lab while working, regardless of their age. However, it is important that such students formulate a long-term plan before commencing their studies.

Many teachers bursting with a research-orientated mentality and a well-equipped research environment



●What are some of Tohoku University’s attractions?

Tohoku University has many faculty members bursting with a research-orientated mentality and playing active roles both in Japan and nationally, in addition to an excelled research environment equipped with a full range of facilities, machines, and materials. The city of Sendai, including the area surrounding the University, is compact and in close proximity to nature, providing an excellent environment for a research lifestyle.

  • 写真1図1
  • 写真2図2
  • 写真3図3



One of my hobbies is photography and cameras. This photograph won a prize in a contest and was used on the cover of a catalog . Just as my job is conducting research related to imaging information, I am interested in various imaging and related devices, not just those used in the medical field. I have also been involved in image file format analysis programs. The imaging for my hobbies and research work are completely different, so I do not choose my research theme based on my interests—that was simply a coincidence. Looking back now, though, I think that there may have been some kind of influence. At the very least, my knowledge of my hobbies is helpful to my research and vice-versa, so I am convinced that basing learning or inquiry on your interests is not a waste of time.


Health Sciences
画像情報学分野 教授
Medical Image Computation Professor
増谷 佳孝
Yoshitaka Masutani


Born in Tokushima Prefecture in 1968. Received a Master’s Degree in Department of Precision Engineering, School of Engineering, The University of Tokyo in 1997. After obtaining his Ph.D. (Engineering), gained experience as a postdoctoral fellow at the Universität Hamburg (Germany) and the University of Chicago (USA). Became a lecturer at the Department of Radiology, The University of Tokyo Hospital in 2001 and a professor at the Department of Biomedical Information Sciences, Faculty of Information Sciences, Hiroshima City University in 2014. Appointed to his current position in 2022. Obtained his Doctor of Medicine degree in 2010, while working at the University of Tokyo Hospital.