CAJ | 학술논문

EPIC ECG 穿戴装置是一种基于新型电势测量元件的非接触电机的穿戴心电获取装置。首先提出了穿戴式心电获取装置整体架构；其次基于心拍的数值表征参数和形态，在已有的支持向量机算法对心电信号进行预测分类的基础上，重点研究如何利用专家库和轮廓模式识别结合的综合特征向量提取方法对分类准确度进行改进；最后利用 MIT-BIH 数据库数据进行了实验验证。实验结果表明，综合特征向量的应用提高了 SVM 的分类准确度。
EPIC ECG 천대장치시일충기우신형전세측량원건적비접촉전궤적천대심전획취장치。수선제출료천대식심전획취장치정체가구；기차기우심박적수치표정삼수화형태，재이유적지지향량궤산법대심전신호진행예측분류적기출상，중점연구여하이용전가고화륜곽모식식별결합적종합특정향량제취방법대분류준학도진행개진；최후이용 MIT-BIH 수거고수거진행료실험험증。실험결과표명，종합특정향량적응용제고료 SVM 적분류준학도。
The real-time cardiac alert algorithm for EPIC ECG wearable device is focused on for study. Firstly , the architecture of wearable ECG acquisition device is proposed overall. Then SVM algorithm is used to analyze and predict EGC signal , mainly focusing on the improvement of classification accuracy by using comprehensive feature vector extraction , which is combined with expert database and profile pattern recognition. Finally, the data from MIT-BIH database is used to conduct verification experiments. Experiment results demonstrate that compared with single feature extraction method , comprehensive feature vector applications improve classification accuracy.