ECG Analysis Using DWT and Wavelet Coefficient to Reduce the Feature and SVM-ICP for Classification and Matching


The Electrocardiogram (ECG) considered as one of the important issue in the medical field (hospitals and clinics), which is used to represent the health of a heart. Increasing patients of heart has supposed to design an automatic computerization technique to classify various abnormalities of the heart activities; to reduce the analysis time and detection mistakes. This research focusing on achieve high performance of classifying abnormal ECG by applying different methods. The first method is Discrete Wavelet Transform (DWT) with 4-level to transform the ECG signal and extract the feature extraction and Wavelet Energy (WE) during feature extraction as feature vector. In classification phase has used Support Vector Machine (SVM) to train datasets and classify the test samples, in matching phase, find closest vector of test to the training datasets method has used by applying Iterative Closest Point (ICP).


ECG, DWT, Wavelet Energy, SVM, ICP