Overcomplete Multi-Scale Dictionaries for Efficient Representation of ECG Signals

Abstract

The electrocardiogram (ECG) was the ?rst biomedical signal where digital signal processing techniques were extensively applied. The ECG is a sparse signal, composed of relevant activations, periods of inactivity, noise and interferences. In this work, we describe an efficient method to construct overcomplete and multi-scale dictionaries for sparse ECG representation using waveforms recorded from real-world patients. Unlike most existing methods, the proposed approach learns the dictionary first, and then applies an efficient sparse inference algorithm to model the signal using the constructed dictionary. As a result, our method is able to deal with long recordings from multiple patients. Simulations on real-world records from Physionet's PTB database show the good performance of the proposed approach.