Automatic continuous ECG monitoring system for over-drug detection in Brugada Syndrome.

This paper is concerned with the automatic control of drug administration in patients suffering from Brugada Syndrome (BS). Drugs such as flecainide, procainamide, ajmaline and pilsicainide should be administrated under carefully controlled electrocardiogram (ECG) monitoring given that the treatment must be stopped if some ECG disturbing conditions appear. These conditions are, among others the development of premature ventricular contraction (PVC), atrial fibrillation (AF) and the widening of the QRS wave. The proposed system can detect these abnormalities by using a pattern recognition approach based on Hidden Markov Models (HMM) with features extracted from three scales of the Wavelet Transform (WT). Performances higher than 98% were reached regarding the classification of normal and abnormal pulses. The system was trained and tested mainly in data from the standard MIT-BIH arrhythmia database.

ECG Data-Acquisition and classification system by using wavelet-domain Hidden Markov Models.

This article is concerned with the classification of ECG pulses by using state of the art Continuous Density Hidden Markov Models (CDHMM's). The ECG signal is simultaneously observed at three different level of focus by means of the Wavelet Transform (WT). The types of beat being selected are normal (N), premature ventricular contraction (V) which is often precursor of ventricular arrhythmia, two of the most common class of supra-ventricular arrhythmia (S), named atrial fibrillation (AF), atrial flutter (AFL), and normal rhythm (N). Both MLII and V1 derivations are used. Run time classification errors can be detected at the decoding stage if the classification of each derivation is different. These pulses are selected for a posterior physician analysis. Experimental results were obtained in real data from MIT-BIH Arrhythmia Database and also in data acquired from a developed low-cost Data-Acquisition System.

Cardiac arrhythmia classification using wavelets and Hidden Markov Models - a comparative approach.

This paper reports a comparative study of feature extraction methods regarding cardiac arrhythmia classification, using state of the art Hidden Markov Models. The types of beat being selected are normal (N), premature ventricular contraction (V) which is often precursor of ventricular arrhythmia, two of the most common class of supra-ventricular arrhythmia (S), named atrial fibrillation (AF), atrial flutter (AFL), and normal rhythm (N). The considered feature extraction methods are the standard linear segmentation and wavelet based feature extraction. The followed approach regarding wavelets was to observe simultaneously the signal at different scales, which means with different level of focus. Experimental results are obtained in real data from MIT-BIH Arrhythmia Database and show that wavelet transform outperforms the conventional standard linear segmentation.