[1L4-OS-18a] 生体信号を活用した医療・ヘルスケアAI

[1L4-OS-18a-02] Novel Method of Blood Pressure Measurement with High Noise Resistance

〇Hirotaka Wada¹, Shihori Uchida¹, Yuki Sasaki¹, Masayuki Koizumi¹, Yukiko Yanagawa¹, Hiroshi Nakajima¹ (1. OMRON Corporation)

[[Online]]

Keywords:blood pressure estimation, noise resistance, machine learning

The cardiovascular events such as heart attacks and strokes are main cause of the mortality in Japan and the other developed countries. Hypertension treatments with blood pressure control is crucial to prevent the disease. Recently, some wearable sphygmomanometers are invented to measure the casual blood pressures. They are small and easy to use for hypertension patients and candidates to control their blood pressure. In order to keep the measurement accuracy of the small device with body movements, we develop the novel method to improve accuracy of blood pressure measurement by the oscillometric-based sphygmomanometer. We select new morphologic feature variables and conducted the experiment to prove the effectiveness of the method. The comparison study on the developed method and the conventional one was done on the dataset collected from 85 subjects by cross-validation. The result showed that the method improves 1.177 of SDE(standard deviation of error) on SBP and 0.657 DBP.