用于室颤节律辨识的多参数融合BP神经网络设计  

Design of BP neural network based on multi-parametes for VF detection

作　　者：余明[1] 陈锋[1] 张广[1] 顾彪[1] 李良喆 王春晨[1] 王丹[1] 吴太虎[1] 

机构地区：[1]军事医学科学院卫生装备研究所,天津300161

出　　处：《军事医学》2016年第10期829-832,838,共5页Military Medical Sciences

基　　金：国家自然科学基金资助项目(81501551)

摘　　要：目的构建一个用于区分室颤节律和非室颤节律的多参数融合BP神经网络。方法从心电信号中提取出18个特征值,分别标记心电信号的形态分布、高斯性、幅度谱、变异性、复杂度等各方面特征;以这些特征值作为输入向量,构建一个多参数融合BP神经网络进行训练,得到一个用于区分室颤节律与非室颤节律的分类器。结果与结论将构建的BP神经网络分别基于VFDB数据库和CUDB数据库进行实验,辨识准确率分别高达98.61%和95.37%;相较于现有方法,辨识性能显著提高。Objective To develop a BP neural network to differentiate between ventricular fibrillation（ VF） and non-VF rhythms. Methods Eighteen metrics were extracted from the ECG signals. Each of these metrics respectively characterized each aspect of the signals,such as morphology,gaussianity,spectra,variability,and complexity. These metrics were regarded as the input vector of the BP neural network. After training,a classifier used for VF and non-VF rhythm classification was obtained. Results and Conclusion The constructed BP neural network was tested with the databases of VFDB and CUDB,and the accuracy was 98. 61% and 95. 37%,respectively.

关 键 词：心电图 室颤 BP神经网络 多参数融合辨识 

分 类 号：R541.76[医药卫生—心血管疾病] TN911.6[医药卫生—内科学]