Font Size: a A A

Automatic Patten Recognition Of ECG Signals Based On Independent Component Analysis Feature Extraction

Posted on:2008-12-31Degree:MasterType:Thesis
Country:ChinaCandidate:X JiangFull Text:PDF
GTID:2178360212476042Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Cardiovascular disease is the leading cause of death around world. Electrocardiogram (ECG) supervising is the most important and efficient way of preventing heart attack. The types of ECG arrhythmia is an important standard to measure the variation of cardiovascular activity. Recognizing arrhythmia on its early stage is very important for diagnosing and forecasting the state of illness, and it is also critical for the further treatment of the patients. Those former research on recognizing heart beat arrhythmia mainly used the ECG morphology and heartbeat interval, wavelet transform or Hermit representation as the feature extraction methods and achieved certain results. This thesis proposes a new feature extraction method based on Independent Component Analysis (ICA) to classify the heartbeat arrhythmia. The method is used to recognize normal beats and the other 13 types of arrhythmia heart beats from the data provided by MIT-BIH arrhythmia database.ICA is a signal processing technique to extract independent components from muti-dimensional mixed signals. It has been widely used in many fields, such as biomedical signal processing, speech recognition, and antenna array processing. This thesis proposes two architectures for feature extraction based on independent component analysis: one is to learn the independent feature basis functions of ECG heartbeat signals, and the other is to learn the independent feature representation of ECG heartbeat signals. From computer experiments, it is demonstrated that both two architectures have their advantages and both are effective for the heartbeat recognition.In addition, an overcomplete feature set which is constructed by different types of feature extraction methods is proposed in the thesis. By combining the characteristics of different types of features, the overcomplete features makes those different features work together and obtains a better feature representation of the ECG signals. In this thesis, an overcomplete feature extraction method combining ICA basis function' s coefficients and wavelet transform coefficients is propose for ECG recognition. Because there are some features that...
Keywords/Search Tags:Independent Component Analysis, ECG, arrhythmia, Principal Component Analysis, Wavelet Transform, Overcomplete Feature, Mutual Information, MIT-BIH arrhythmia database
PDF Full Text Request
Related items