Development and Validation of an Algorithm for the Digitization of ECG Paper Images

Area of Science:

• Biomedical Engineering
• Medical Informatics
• Cardiology

Background:

• Paper-based electrocardiograms (ECGs) are common but prone to deterioration, hindering long-term patient health assessments.
• Digitizing ECGs is crucial for applying engineering signal processing techniques and ensuring data integrity.
• Existing methods for ECG digitization may lack efficiency or accuracy, necessitating improved automated solutions.

Purpose of the Study:

• To develop an automatic algorithm for converting paper-based ECG images into digital ECG signals.
• To validate the accuracy and reliability of the developed digitization algorithm.
• To assess the potential of the algorithm as a tool for cardiologists.

Main Methods:

• An automatic conversion algorithm was designed for digitizing paper ECGs.
• The algorithm was tested on a dataset of 16 subjects, including those with cardiac pathologies.
• Quantitative analysis included evaluating repeatability, reproducibility, and accuracy of digitized ECG time parameters.

Main Results:

• The algorithm demonstrated high efficiency with an average Pearson correlation coefficient of 0.94.
• Measurement errors for key ECG time parameters (e.g., R-R interval, QRS duration) were consistently below 1 mm.
• The digitization method proved accurate for both the entire signal and specific temporal measurements.

Conclusions:

• The developed algorithm effectively converts paper ECGs into accurate digital signals.
• The algorithm's performance suggests its suitability for integration into clinical tools for cardiologists.
• This digitization method enhances the preservation and analysis of historical and current ECG data.