摘要
目的:研究重症监护病人基于心电和血压信号数据融合的心率估计算法,抑制监护仪的错误报警。方法:分别从心电和血压信号计算逐搏心率,通过分析信号特征得出反映信号质量好坏的信号质量指数,应用卡尔曼滤波方法分别对基于心电和血压分析得出的心率进行最佳估计,用卡尔曼滤波的残差和信号质量指数作为权重系数进行心率数据融合。应用融合心率算法对美国麻省理工学院多参数智能重症监护数据库II中记录的监护仪产生的2584次严重心动过缓和严重心动过速报警数据重新分析,以期抑制监护仪的错误报警。结果:本算法对2584次报警数据分析,对真实报警的正确识别率为99.64%,对错误报警的抑制率为70.66%。结论:基于数据融合的心率估计可有效抑制监护仪的错误报警,提高报警的准确率。
Objective: To develop a heart rate (HR) estimation method based on data fusion of electrocardiogram (ECG) and arterial blood pressure (ABP) from intensive care unit (ICU) patients, and to suppress the false alarms oflCU monitors. Methods: Beat-by-beat HR was calculated separately from ECG and ABP. Signal quality indices (SQI) were obtained by analyzing characteristics of each waveform. HR from the ECG and ABP was tracked with a Kalman filter and weighted by the Kalman filter's residual error and SQI to perform the HR data fusion. This method was evaluated using 2584 episodes of extreme bradycardia and extreme tachycardia alarms in the Multi-parameter Intelligent Monitoring for Intensive Care Ⅱ database. Results: Our algorithm detected correctly 99.64% of the true alarms and suppressed 70.66% of the false alarms. Conclusions: The algorithm could suppress false alarms oflCU monitors efficiently and improve the accuracy rate of alarms.
出处
《中国医学物理学杂志》
CSCD
2008年第3期676-678,共3页
Chinese Journal of Medical Physics
关键词
心率估计
数据融合
重症监护
错误报警
heart rate estimation
data fusion
ICU monitor
false alarm