Predicting the occurrence of coughing and taking measures against splash ― Tohoku University develops a system to suppress the spread of respiratory infections

Respiratory tract infections, such as the novel coronavirus infection, are thought to spread via cough droplets. As a result of this, the use of air curtains for protection while coughing may make it possible to control the spread of infection. A research group including Professor Yasuyuki Shiraishi, Professor Tomoyuki Yambe and Assistant Professor Akihiro Yamada of the Institute of Development, Ageing, and Cancer, Tohoku University, Associate Professor Norihiro Sugita and Professor Makoto Yoshizawa of the Graduate School of Engineering and Graduate School of Biomedical Engineering, Tohoku University, focused on droplet transmission, and developed a system capable of "diagnosing" and "predicting" the "occurrence of coughing" from 3D images of the body surface. AI analysis of the time series of the "exhalation phase" using the body surface image has made three-dimensional surface reconstruction and motion analysis of the human body even from images that originally contained only two-dimensional information possible, thereby facilitating quantitative diagnosis and prediction of the occurrence of a "cough" by AI. This was announced on November 1st at IEEE EMBC, the world's largest academic society for electronic information systems, held online in Guadalajara, Mexico.

Currently, Associate Professor Hitomi Anzai and Professor Makoto Ohta of the Institute of Fluid Science, Tohoku University are conducting research on fluid simulation of droplets. Using the simulation results obtained from supercomputers, the speed and distribution of droplet transmission can be determined, and methods of blocking transmission can be predictively designed. For example, if the occurrence of a "cough" is predicted and an air curtain or something similar is deployed only in front of the person coughing to create a system that prevents droplets from being transmitted, the occurrence of unpredictable clusters in hospitals, etc., can be greatly reduced. This system could also be deployed in environments such as schools, restaurants, and transportation vehicles and facilities.