Visualization of physiologic signals based on Hjorth parameters and Gramian Angular Fields

Visual analysis is an important part of interactive data exploration. Visualization is especially important for analysis of physiological signals derived from sensors placed on the human body or implanted. Biomedical signals can provide valuable information about the state of a subject, but usually they are difficult to interpret due to large amount, high complexity and high-dimensionality. Here we present a novel method for two-dimensional visualization of physiological time series based on Gramian Angular Field (GAF) representations of Hjorth parameters of time series. Illustrative examples for electrocardiography (ECG) signals are presented. The practical use of the visualization method is demonstrated on the arrhythmia case classification task. We report 86% accuracy of classification (using KNN) of Premature Ventricular Contraction (PVC) beats vs. normal beats using the statistical features extracted from images generated using the proposed method.