复杂生理信号研究资源-PhysioNet及其在抑制监护仪错误报警中的应用开发

PhysioNet-A Resource of Complex Physiological Signals and the Application in Depressing False Alams of Intensive Care Monitors

目的:研究复杂生理信号研究资源-PhysioNet的应用开发方法,研究抑制监护仪错误报警的算法。方法:首先详细介绍了PhysioNet以及它的三个相互关联的组成部分:数据库PhysioBank,软件库PhysioToolkit和网络资源平台PhysioNet;其次,研究了应用PhysioNet进行算法开发研究的方法,然后研究了抑制监护仪错误报警的算法,利用PhysioToolkit软件包开发程序,取PhysioBank里的MIMIC II数据库中的心动过缓和心动过速报警数据进行分析,分类检测了监护仪产生的报警信息,对算法的性能进行了整体评价。结果:算法对真实报警的正确识别率为99.64%,对错误报警的抑制率为66.18%。结论:PhysioNet是分析人体复杂生理信号的重要的数据资源和研究开发平台,我们将其应用开发,对抑制监护仪错误报警算法的性能进行了整体评价,取得了较好的效果。

Objective： To study the methods of application and development using the Physionet, which is a significant research resource of complex physiological signals, and to study an algorithm for suppressing the false alarms produced by monitors. Methods： Firstly, We introduced the PhysioNet in detail, and its three interdependent components： PhysioBank （database）, PhysioToolkit （library of software） and PhysioNet （net platform for resource）. Secondly, we studied the methods of algorithm development using the Physionet. Then we studied an algorithm for suppressing the false alarms produced by monitors. The algorithm was programmed using PhysioToolkit. We used the Multi-Parameter Intelligent Monitoring for Intensive Care （MIMIC） II database in PhysioBank, which corresponded to tachycardia and bradycardia alarms produced by the monitors, to evaluate our algorithm as a whole with detecting these classified alarms. Results： The correct identifying rate of true alarms using the algorithm is 99.64%, and the suppression rate of false alarms is 66.18%. Conclusions： As a vital data resource and also a net platform for research and application in analyzing complex physiological signals, PhysioNet was used and developed to evaluate the algorithm for suppressing the false alarms produced by monitors, and exciting results arise.