A toolkit for synthesis, analysis, and digitization of electrocardiogram images
The ECG (Electrocardiogram) is a widely used and accurate diagnostic tool for cardiovascular diseases. Traditionally recorded in printed formats, the digitization of ECGs holds immense potential for training machine learning (ML) and deep learning models in algorithmic ECG diagnosis. However, physical ECG archives are at risk of deterioration. Simply scanning printed ECGs is insufficient, as ML and deep models require time-series data from ECGs. This necessitates the digitization and conversion of paper ECG archives into time-series data.
To address these challenges, we have developed methods and tools for generating synthetic ECG images on standard paper-like ECG backgrounds, faithfully replicating realistic artifacts. Our approach includes plotting ECG time-series data on ECG paper-like backgrounds, followed by applying various distortions such as handwritten text artifacts, wrinkles, creases, and perspective transforms to the generated images while ensuring the absence of personally identifiable information. These synthetic ECG images provide a valuable resource for developing and evaluating ML and deep learning models for ECG digitization and analysis.
In this repository, you will find tools and utilities for analyzing scanned ECG images, generating realistic synthetic ECG images, and digitizing them into time-series data. Our aim is to facilitate the digitization and analysis of ECG archives, ultimately enabling the advancement of computerized ECG diagnosis.
Contributions and feedback are welcome and encouraged from our user community as we strive to revolutionize the digitization and analysis of ECG data using state-of-the-art deep learning and image processing techniques.
Scanning ECG images and their digitization have some fundamental limitations and requirements rooted in signal and image processing theory. A brief tutorial on these requirements is provided in a short note here.
Below are the toolsets for the forward and inverse processes of ECG image generation and digitization:
Please include references to the following articles in any publications:
-
Kshama Kodthalu Shivashankara, Deepanshi, Afagh Mehri Shervedani, Matthew A. Reyna, Gari D. Clifford, Reza Sameni (2024). ECG-image-kit: a synthetic image generation toolbox to facilitate deep learning-based electrocardiogram digitization. In Physiological Measurement. IOP Publishing. doi: 10.1088/1361-6579/ad4954
-
ECG-Image-Kit: A Toolkit for Synthesis, Analysis, and Digitization of Electrocardiogram Images, (2024). URL: https://github.com/alphanumericslab/ecg-image-kit
- Deepanshi, Department of Biomedical Informatics, Emory University, GA, US
- Kshama Kodthalu Shivashankara, School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, US
- Matthew A Reyna, Department of Biomedical Informatics, Emory University, GA, US
- Gari D Clifford, Department of Biomedical Informatics and Biomedical Engineering, Emory University and Georgia Tech, GA, US
- Reza Sameni (contact person), Department of Biomedical Informatics and Biomedical Engineering, Emory University and Georgia Tech, GA, US
Please direct any inquiries, bug reports or requests for joining the team to: ecg-image-kit@dbmi.emory.edu.
