High Precision Digitization of Paper-Based ECG Records: A Step Toward Machine Learning

Abstract

Introduction: The electrocardiogram (ECG) plays an important role in the diagnosis of heart diseases. However, most patterns of diseases are based on old datasets and stepwise algorithms that provide limited accuracy. Improving diagnostic accuracy of the ECG can be done by applying machine learning algorithms. This requires taking existing scanned or printed ECGs of old cohorts and transforming the ECG signal to the raw digital (time (milliseconds), voltage (millivolts)) form. Objectives: We present a MATLABbased tool and algorithm that converts a printed or scanned format of the ECG into a digitized ECG signal. Methods: 30 ECG scanned curves are utilized in our study. An image processing method is first implemented for detecting the ECG regions of interest and extracting the ECG signals. It is followed by serial steps that digitize and validate the results. Results: The validation demonstrates very high correlation values of several standard ECG parameters: PR interval 0.984 +/− 0.021 (p-value < 0.001), QRS interval 1+/− SD (p-value < 0.001), QT interval 0.981 +/− 0.023 p-value <0.001, and RR interval 1 +/− 0.001 p-value <0.001. Conclusion: Digitized ECG signals from existing paper or scanned ECGs can be obtained with more than 95% of precision. This makes it possible to utilize historic ECG signals in machine learning algorithms to identify patterns of heart diseases and aid in the diagnostic and prognostic evaluation of patients with cardiovascular disease. INDEX TERMS Electrocardiogram, digitization, Matlab tool, image processing.

Classification of COVID-19 electrocardiograms by using hexaxial feature mapping and deep learning

“…This background is removed within two steps. In the first step, the background lines were removed by filtering the input densities with a density mapping function [ 57 ], because the background has denser or softer RGB values than the curves expressing the ECG signal. This is essentially a contrast enhancement process.…”

Digitizing ECG image: new fully automated method and open-source software code

“…Several other algorithms for paper ECG digitization have been previously developed. [6,7,[30][31][32][33][34][35][36][37][38] Unlike several earlier digitizing applications that scaled-down image resolution to reduce the computational cost, [7,31] we elected to use high-resolution images (600 dpi), and output All rights reserved. No reuse allowed without permission.…”

“…Several other algorithms for paper ECG digitization have been previously developed. [6,7,[30][31][32][33][34][35][36][37][38] Unlike several earlier digitizing applications that scaled-down image resolution to reduce the computational cost, [7,31] we elected to use high-resolution images (600 dpi), and output high-resolution ECG signal, taking advantage of computationally-efficient Python libraries. [18] Furthermore, we preserved the high resolution of the digitized ECG signal, providing users an opportunity to implement their preferred subsequent ECG signal processing approach.…”

Combining Optical Character Recognition With Paper ECG Digitization

“…While previous studies utilize a connected components based approach to extract ECG lead segments [14] , [16] , in the current work we employ CCA along with a distance-based metric to allocate the connected components to the appropriate ECG row waveforms. Although automatic segmentation strategies for ECG waveforms exist [16] , these strategies are restricted because they assume rigid placement of signals from various ECG leads within the paper record. The region containing each waveform is detected based on bounding boxes, which may lead to inaccurate digitization in instances wherein the signals overlap with one another.…”

“…The second barrier is the lack of open accessible software for digitization of ECG records. This is hampered by the lack of generalizability across vendors of digital ECG systems [16] , [19] , and limited availability to researchers [19] , [20] due to their proprietary nature. While validated open source applications exist for extraction of ECG waveforms from ECG records stored in PDF format [21] , [22] , these applications do not extend to digitization of scanned stored images.…”