We present an analysis method of electrocardiogram (ECG) to determine patients' ventricular late potentials (VLPs) characteristics by means of wavelet packet transform and entropy (ApEn). The results indicate that the patients with VLPs tend to have higher ApEn values than do those without. So ApEn can be used as an indicator and a BP neural network for distinguishing between the two patient groups. The experiment result shows that this method has higher distinguishability.
Electrocardiography ; methods ; Entropy ; Heart Ventricles ; physiopathology ; Humans ; Neural Networks (Computer) ; Signal Processing, Computer-Assisted ; Ventricular Function