Proposal of Feature Extraction from Wavelet Packets Decomposition of QRS Complex for Normal and Ventricular ECG Beats Classification

Long term holter monitoring is widely applied to patients with heart problems such as arrhythmias. The primary task of computer aided systems in holter ECG evaluation is to distinguish between different beat types. In this work we investigate the use of wavelet packets transform as a mean to extract features capable of providing the information needed by a classifier for discrimination between Ventricular (V) and Normal beats (N). We designed approach for feature extraction based on wavelet packets and template matching. We used MIT database for test of this approach. The database was divided into to subsets (testing and validation) and we obtained sensitivity 97,7% and 91,9%, specificity 95,1% and 87,1%, overall accuracy 96,3% and 90,4% on the first subset and second subset respectively.