心电信号的房颤自动识别算法

为提高房颤识别的准确率和效率,提出基于心电信号的房颤自动识别算法.首先,采用二阶差分阈值法对R波进行初筛;然后,引入指数移动平均的方法动态验证初选出的R波,通过整体上移心电图、结合一阶差分阈值和实时调整窗口时间的方法解决影响R波准确率的问题;最后,基于临床经验,提出3条判别规则,从而筛选出相应的房颤片段.实验结果表明:文中算法的阴性预测率和特异性分别可达99.7%和99.8%,可有效帮助医生减少阅读量,提高工作效率和诊断准确率.

Automatic Recognition Algorithm of Atrial Fibrillation Based on ECG

In order to improve the accuracy and efficiency of atrial fibrillation recognition, an automatic atrial fibrillation algorithm based on an electrocardiosignal is proposed. Firstly, the second-order difference threshold method is used to preliminary screening the R-wave. Then, the method of exponential moving average is introduced to verify the initial R-wave dynamically. The problem affecting the accuracy of R-wave is solved by moving the electrocardiograph(ECG)as a whole, combining the first-order differential threshold and adjusting the real-time window. Finally, based on clinical experience, three criteria are proposed to screen out the corresponding atrial fibrillation fragments. The experimental results show that the negative predictive rate and specificity of the proposed algorithm can reach 99.7% and 99.8% respectively, which can effectively help doctors reduce the amount of reading, improve work efficiency and diagnostic accuracy.