A Fusion Method Of ECG Signal Identification

In the social and economic rapid development today, there are a lot of fake crime has seriously impacted the safety of information, so the application of the identification technology has been widely concern. And based on biometric identification technology is mainly through the verification of the biological characteristics of human itself, the characteristics of ecg signals recognition is obtained by collecting living of bioelectricity, and complexity caused by itself is not easy to fake and forged, therefore, based on the identity of the ecg signal recognition technology has a broader scope. And this paper mainly studied the recognition algorithm, on the basis of eliminating noise and interference of the ECG signal, determined the cycle of ECG signal, again using frequency spectrum feature extraction and feature extraction of time-frequency joint analysis of the two feature extraction method, the signal is analyzed and validated, based on two methods of feature extraction was put forward after the identification results of decision-making fusion scheme. The paper also introduces the heart sound signal acquisition and analysis, this paper work is as follows:First of all, this paper first introduces the mechanism of electrocar diogram and heart electric signal waveform, after analysis of the noise in ecg signal, on the basis of this, put forward to eliminate the noise and interference in ECG signals with wavelet transformation, again through the determination of the ecg signal peak point positioning out his one cardiac cycle for processing.Secondly, proposed the ecg signal amplitude spectrum feature extraction and feature extraction of time-frequency joint analysis of the two feature extraction methods. First studied the basic principle and algorithm step of each algorithm, and then through the MIT-BIH database, each in vector maximum error recognition algorithm compares the two methods of feature extraction research.Finally, through the comparative study of the identification results of two kinds of algorithm, puts forward the decision level fusion scheme for extracting method based on the features of two kinds of recognition results, with MIT/BIH database data is verified. Also introduces the heart sound signal acquisition and analysis.