Development of a non-contact infection and heatstroke screening system adopting neural network

2018 Fiscal Year Final Research Report

• Project/Area Number: 16K01365
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Biomedical engineering/Biomaterial science and engineering
• Research Institution: Tokyo Metropolitan University
• Principal Investigator: Matsui Takemi 首都大学東京, システムデザイン研究科, 教授 (50404934)
• Co-Investigator(Kenkyū-buntansha): 孫 光鎬 電気通信大学, 大学院情報理工学研究科, 助教 (80756677)
• Project Period (FY): 2016-04-01 – 2019-03-31
• Keywords: 生体医工学 / 感染症 / 熱中症 / マイクロ波レーダ

Outline of Final Research Achievements

We conducted a research aiming to develop non-contact infection screening system and executed its performance evaluation.
We developed a portable and a desktop type systems which measure respiration rate, body temperature and heart rate. By pressing photoplethysmographic sensor of the portable system to patient's carotid artery, it measures heart rate, its integrated infrared thermometer measures body surface temperature, and 24GHz Doppler radar measures respiration rate. We evaluated the performances at general clinic and elderly care type hospital. Our system could measure the vital signs within 10 seconds and discriminated infected and non-infected using machine learning algorithm. Moreover, we enabled automatic screening after we conducted clinical trials with influenza patients and pediatric pneumonia patients in Mongolia and Japan.

Free Research Field

生体医工学

Academic Significance and Societal Importance of the Research Achievements

本感染症スクリーニングシステムでは、インフルエンザ患者の呼吸数、心拍数、体温を用いて機械学習を行い、ランダムツリーアルゴリズムを採用した。季節性インフルエンザ患者を対象とする試験臨床を行い約96％の感度と特異度を達成し、(Journal of Infection 2019, Impact Factor=4.6)インフルエンザの簡易検査キットよりも高い感度、特異度で感染症のスクリーニングが可能となった。
本システムは、感染症の流行の発生をいち早く把握できるほか、検温時に入院患者をモニタリングすることで、重症化へ至る前兆を把握する事が可能である。