Using supervised fuzzy clustering and CWT for ventricular late potentials (VLP) detection in high-resolution ECG signal

Abstract

Ventricular Late Potentials (VLPs) are lowamplitude, high-frequency signals that appear at the end of the QRS complex of a High-Resolution ECG (HRECG) records. VLPs are clinically useful for identifying post-MI (Myocardial Infarction) patients prone to Ventricular Tachycardia (VT) and Sudden Cardiac Death (SCD). In this paper, the Continuous Wavelet Transform (CWT) and a supervised fuzzy clustering algorithm are used together to detect VLPs. The terminal part of the QRS complex in the Vector Magnitude (VM) waveform is processed with the CWT to extract a feature vector. Resulting time-scale representation is subdivided into several sub bands, and the sum of the squared decomposition coefficients is computed in each region. Finally, a supervised Fuzzy clustering method, trained by an appropriate set of these feature vectors, is applied to this data in order to identify VLP.