The Power of Tensor-Based Approaches in Cardiac Applications

Abstract

The electrocardiogram (ECG) is a biomedical signal that is widely used to monitor the heart and diagnose cardiac problems. Depending on the clinical need, the ECG is recorded with one or multiple leads (or channels) from different body locations. The signals from different ECG leads represent the cardiac activity in different spatial directions and are thus complementary to each other. In traditional methods, the ECG signal is represented as a vector or a matrix and processed to analyze temporal information. When multiple leads are present, most methods process each lead individually and combine decisions from all leads in a later stage. While this approach is popular , it fails to exploit the structural information captured by the different leads. Recently, there is a trend towards the use of tensor-based methods in biomedical signal processing. These methods represent the signals by ten-sors, which are higher-order generalizations of vectors and matrices that allow the analysis of multiple modes simultaneously. In the past years, tensor decomposition methods have been applied on ECG signals to solve different clinical challenges. This chapter discusses the power of different tensor de-compositions with a focus on typical ECG problems that can be solved using tensors.