摘要
提出了一种基于主成分分析法对ECG信号进行特征表述的身份识别新方法。在预处理阶段对ECG信号进行滤波、分段、归一化、抽样处理,然后计算ECG信号的协方差矩阵及协方差矩阵的特征值和特征向量,其中较大的特征值对应的特征向量具有与ECG相似的形状,利用这些特征向量可描述、表达和逼近ECG信号并用于后续的身份识别。实验结果表明:相对于ECG基点特征提取技术,该方法提高了录用率,获得了较好的识别效果。
This paper presents eigen ECG, which is based on principal component analysis, a new method for hu- man identification. In the pre-processing stage,the ECG signal was filtered, segmented, normalized and sampled. Then the covariance matrix of the ECG signal, and its eigenvalues and eigenvectors were calculated. The corre- sponding eigenvectors of the larger eigenvalues are similar to ECG in shape. They can be used for the descrip- tion, expression and approximation of ECG signal and the subsequent identification. Experimental results show that compared with ECG fiducial extraction, this method improves the enrolling rate and takes a better recogni- tion effect.
出处
《苏州科技学院学报（自然科学版）》
CAS
2010年第3期55-58,共4页
Journal of Suzhou University of Science and Technology （Natural Science Edition）
基金
山东省自然科学基金资助项目(y2006G03
y2007G14)
关键词
主成分分析
特征提取
归一化
身份识别
principal component analysis
feature extraction
normalization
human identification
