Title :
ECG identification based on Matching Pursuit
Author :
Zhao, Zhidong ; Yang, Lei
Author_Institution :
Coll. of Electron. & Inf., Hang Zhou DianZi Univ., Hangzhou, China
Abstract :
Electrocardiogram (ECG) reflects cardiac electrical activity, and varies from person to person, which could be used for biometrics identification. ECG biometrics identification algorithm is presented in this paper based on Matching Pursuit (MP) and Support Vector Machine (SVM). The ECG signal is decomposed into atoms by sparse decomposition with Gabor dictionary which contain ECG signal´s important information. SVM is used to identify the person. Experiment shows that the performance of the system over 20 subjects is 95.3%.
Keywords :
Gabor filters; bioelectric potentials; biometrics (access control); decomposition; electrocardiography; medical signal processing; support vector machines; ECG biometrics identification algorithm; ECG identification; ECG signal decomposition; Gabor dictionary; SVM; cardiac electrical activity; electrocardiogram; matching pursuit; sparse decomposition; support vector machine; Biometrics; Databases; Electrocardiography; Feature extraction; Heart beat; Matching pursuit algorithms; Support vector machines; biometric identification; electrocardiogram; matching pursuit; support vector machine;
Conference_Titel :
Biomedical Engineering and Informatics (BMEI), 2011 4th International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4244-9351-7
DOI :
10.1109/BMEI.2011.6098470
