Abstract: An algorithm to extract features of electro-cardio-gram (ECG) signal based on wavelet multi-resolution analysis is developed. The wavelet transform is used for the segmentation and the feature extraction of ordinary 12-lead ECG. The support vector machine is used to classify the unknown ECG signal. The result shows that feature extraction technology has an optimistic prospect.

Key words: feature extraction, wavelet transform, support vector machine, ECG classification, segmentation, feature extraction, wavelet transform, support vector machine, ECG classification