一种实用的可除颤节律识别算法

A Practical Algorithm for Defibrillation Rhythm Detection

根据可除颤节律的特点,一种基于多特征参数联合判断的可除颤节律自动识别算法被提出。首先,对心电信号进行预处理;然后分别计算信号的峰值斜率、幅度概率密度、波峰幅度均方差、波形复杂度和心率参数;再对这些特征参数进行综合决策,从而实现可除颤节律的自动识别。采用公共标准数据库(CU、AHA、MIT-BIH、VFDB、NST)和迈瑞公司临床数据库对算法进行评测,评测结果表明该算法灵敏度和特异度都达到了较高水平,充分满足临床应用的安全有效性要求。

According to the characteristics of defibrillation rhythm,an algorithm for automatic defibrillation rhythm detection was proposed base on multiple characteristic parameters.First,the original electrocardiosignal preprocessed.Then,the peak slope,amplitude probability density,mean square deviation of peak amplitudes,waveform complexity,and heart rate parameters of the signal were calculated,respectively.Finally,the defibrillation rhythm was automatically detected by comprehensively analyzing these parameters through a decision tree method.The algorithm was evaluated using public standard databases,such as CU,AHA,MIT-BIH,VFDB,and NST,and Mindray clinical database.The results indicate that the proposed algorithm shows a good performance in both sensitivity and specificity,and is safe and effective for clinical application.