Font size

  • S
  • M
  • L

Laboratory of Media Dynamics
Graduate School of Information Science and Technology
Hokkaido University

Collaboration with Medical Department

Improving Stomach Cancer Risk Evaluations through Collaboration between Medicine and Information Science

In 1994, the World Health Organization announced that H. Pylori infection is a cause of cancer risk. This knowledge has made its way to Japan, where medical professionals began to focus on reducing cancer risk through disinfection, and medical facilities such as hospitals implemented tests for H. pylori infection. In order to improve the correctness of the stomach cancer risk evaluations, it is important to include several different tests, such as blood tests and the analysis of stomach images obtained via X-ray and endoscopy. However, since that involves increased labor for the medical professionals and increased cost, computer systems that can automatically perform the image analysis are in demand.

The Laboratory of Media Dynamics studies “Multimedia Signal Processing Theory”, which enables the extraction of latent knowledge through the collaborative processing of data in various forms and media, such as images, video, acoustic signals and sensor data. Furthermore, the study of “Correlation Analysis Theory”, which involves revealing the mutual relationships between different data in various forms, has also yielded important results. The Laboratory of Media Dynamics is developing a computer system for stomach cancer risk evaluation based on these results.

An important problem of conventional examinations was that since they focused on stomach X-ray images and endoscopy images in isolation, they did not allow the images to be analyzed integrally. Our research applies “Multimedia Signal Processing Theory” to improve the correctness of examinations by integrating the results of various forms of observed data, such as X-ray and endoscopy images. Furthermore, we apply “Correlation Analysis Theory” to determine the relationship between the doctor’s past examination history and the observed data, visualize it, and present it to the doctor, conveying new knowledge.

This research extracts information about the mechanism behind stomach cancer risk, and reveals it through a collaboration between two different fields, medicine and information science. This revelation is the foundation of the research. Since different fields mutually desire new information that could be understood, this research creates new value, and contributes to the improvement of existing techniques in both fields. We also expect it to go beyond stomach cancer risk evaluations and encourage the creation of new academic knowledge in other fields.