What is the clinical evidence from this study?

Study design: Other. Population: Cardiovascular Disease (n=928). Intervention: Convolutional Neural Network (CNN). Primary outcome: Classification accuracy on test data.

What question did this study set out to answer?

The research aims to develop an automated method for classifying ECG images to enhance cardiovascular disease detection.

June 27, 2026International Journal of Technology & Emerging Research0 citationsOpen Access

Deep Convolutional Network Modeling for ECG Image Analysis in Cardiovascular Disease Detection

PDPilla Divya DPDr. M. V. V. Siva Prasad

Key Result

A custom Convolutional Neural Network achieved a classification accuracy of 97.92% on test data for categorizing ECG images into four cardiovascular conditions.

Key Points

The research aims to develop an automated method for classifying ECG images to enhance cardiovascular disease detection.
Developed a convolutional neural network (CNN) for ECG image analysis.
Utilized a dataset of 928 ECG images in four categories: Normal, Abnormal, History of Myocardial Infarction, and Myocardial Infarction.
Implemented preprocessing steps including grayscale conversion, noise reduction, cropping, resizing, and normalization.
Achieved 97.92% classification accuracy on test data.
Demonstrated strong precision, recall, and F1-scores across all ECG categories.
Deployed a Flask-based web application for real-time predictions and visualizations.

Structured PICO

Population

928 ECG images from a public Kaggle dataset were used to train and evaluate a deep learning model for cardiovascular disease classification.

Intervention

Custom Convolutional Neural Network (CNN) architecture for automated ECG image classification

Outcome

Classification accuracy on test datasurrogate

A custom CNN-based framework demonstrated high accuracy (97.92%) for automated classification of ECG images into normal and MI-related categories.

Limitations

The dataset size was relatively small, and additional large-scale, multi-center datasets would further validate the model's robustness.
Misclassifications between Normal and History of MI cases suggest that subtle waveform variations remain challenging.
The system is intended to be a decision-support tool, not a replacement for professional medical judgment.

Abstract

Cardiovascular diseases (CVDs) remain the foremost cause of global mortality, accounting for nearly one-third of deaths worldwide. Early detection of cardiac abnormalities is essential to reduce mortality and ensure timely treatment. Electrocardiograms (ECGs) are among the most widely used diagnostic tools for monitoring cardiac activity. However, manual interpretation of ECGs is error-prone and highly dependent on medical expertise. This paper presents a Convolutional Neural Network (CNN)-based framework for automated ECG image classification. The dataset, consisting of 928 ECG images across four categories— Normal, Abnormal, History of Myocardial Infarction (HMI), and Myocardial Infarction (MI) was preprocessed through grayscale conversion, noise reduction, cropping, resizing, and normalization. A custom CNN architecture was trained on this dataset, achieving a classification accuracy of 97.92% on test data, with strong precision, recall, and F1-scores across all categories. The system was deployed using a Flask-based web application that provides real-time predictions and visualizations. The proposed solution demonstrates the applicability of deep learning in medical diagnostics, offering a reliable and scalable approach for CVD detection. Keywords: Electrocardiogram (ECG); Convolutional Neural Network (CNN); Cardiovascular Disease Detection; Deep Learning; Medical Image Classification; Flask Web Application

问 AI

Bookmark

View Full Paper