Clustering of ECG segments for patients before sudden cardiac death based on Lagrange descriptors

Abstract

Novel approach to clustering of ECG segments based on Lagrange descriptors is presented in this paper. The approach starts by extracting 2D features with the help of Lagrange descriptors. Then the features are transformed to latent vectors which are clustered using K-means algorithm. The object of the research is to visualize the dynamics of clusters of 2D features of segments of ECG before sudden cardiac death happens to a patient.