A Conversion Algorithm for ECG signals on a 2D array based on Digital Signal Processing

Abstract

This work proposes a computational algorithm to convert digital files containing electrocardiogram (ECG) information into 1D signals. Many medical databases have in storage files containing ECG information that is not easy to process for computational algorithms. Digitization by the proposed method makes it possible to modernize the databases of many health centers in order to perform post-processing of the signals obtained. This method is based on applying digital signal processing techniques to images obtained from a PDF file produced by an electrocardiograph. The proposed algorithm takes into consideration the thickness of the printed signal in the PDF image so that it does not introduce distortion in the final 1D signal. Due to the distribution of the ECG signals on the PDF files the algorithm identifies and segments the signals on 2 dimensions. The results show that the proposed method can correctly reproduce the information of the ECG waves captured in the PDF file regardless of the elements outside the ECG signal such as the background grid or the different information indicators, whether they are labels or references of the ECG signals. The algorithm has an accuracy of 95% based on the statistical analysis performed for all samples.