运用双向长短期记忆模型的心拍分类算法

为提高心拍的分类效果,研究基于双向长短期记忆(BiLSTM)模型的深度学习算法.首先,采用“双斜率”法对心电信号进行预处理;然后,设计自适应阈值对预处理后的心电信号进行QRS波定位,并依据R波波峰分割截取心拍;最后,采用BiLSTM模型的深度学习算法对获取的心拍形态进行分类.使用MIT-BIH心率失常数据库验证算法有效性,实验结果表明:文中算法对正常或束支传导阻滞(N)、室上性异常(S)、心室异常(V)、融合(F)类型的敏感性分别为98.56%,97.10%,93.33%,79.52%,特异性分别为98.38%,98.08%,98.54%,99.65%;与传统的支持向量机等方法相比,文中算法能够进一步提高心拍分类的正确率.

Heartbeat Classification Algorithm Using Bi-Directional Long-Short-Term Memory Model

In order to improve the classification effect of heart beats, a deep learning algorithm based on bi-directional long and short-term memory(BiLSTM)model is studied. Firstly, the “double slope” method is used to preprocess the electrocardiogram signal. Then, an adaptive threshold is designed to perform the preprocessed electrocardiogram signal. QRS waves are located, and heartbeats are intercepted according to R wave peak segmentation. Finally, the deep learning algorithm of BiLSTM model is used to classify the acquired heartbeat shapes. MIT-BIH arrhythmia database is used to verify the effectiveness of the algorithm. The experimental results show that the sensitivity of the proposed algorithm to bundle branch block(N), supraventricular abnormality(S), ventricular abnormality(V), and fusion(F)is 98.56%, 97.10%, 93.33%, 79.52%, specificity were 98.38%, 98.08%, 98.54%, 99.65%, respectively; compared with the traditional support vector machine method, the proposed algorithm can further improve the accuracy of heartbeat classification.