期刊
ARTIFICIAL INTELLIGENCE IN MEDICINE
卷 97, 期 -, 页码 143-151出版社
ELSEVIER SCIENCE BV
DOI: 10.1016/j.artmed.2018.12.005
关键词
Cuffless blood pressure estimation; Photoplethysmogram intensity ratio; Multivariate adaptive regression spline; Pulse transit time; Taken's theorem
Continuous cuffless blood pressure (BP) monitoring has attracted much interest in finding the ideal treatment of diseases and the prevention of premature death. This paper presents a novel dynamical method, based on pulse transit time (PIT) and photoplethysmogram intensity ratio (PIR), for the continuous cuffless BP estimation. By taking the advantages of both the modeling and the prediction approaches, the proposed framework effectively estimates diastolic BP (DBP), mean BP (BP), and systolic BP (SBP). Adding past states of the cardiopulmonary system as well as present states of the cardiac system to our model caused two main improvements. First, high accuracy of the method in the beat to beat BP estimation. Second, notwithstanding noticeable BP changes, the performance of the model is preserved over time. The experimental setup includes comparative studies on a large, standard dataset. Moreover, the proposed method outperformed the most recent and cited algorithms with improved accuracy.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据