Ghostden: Hybrid Framework for ECG Image Based Heart Disease Classification

View at publisher Summarise with VIRAT
Authors: S. Hemalatha, D. Sivabalaselvamani, T. Kavitha, K. Kaliswary, V. Dhanush, J. Jaisapthagiri
Cited by: 0
References: 18

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Heart disease is one of the major causes of death around the world today. Detecting heart disease early through electrocardiogram (ECG) analysis can significantly improve chances for timely medical treatment. This paper introduces GhostDen, a deep-learning-based framework that classifies ECG images into four main classes: Normal, Abnormal, History of Myocardial Infarction (MI) and Myocardial Infarction. The presented system employs GhostNet as a lightweight and efficient base for feature extraction. To cause the model to focus on critical patterns shown in ECG images, two attention mechanisms, Squeeze-and-Excitation (SE) and Convolutional Block Attention Module (CBAM), are integrated into the architecture. Subsequently, dense layers and softmax activation are used to accomplish the classification. Experiment-based results show the SE model achieves a classification accuracy of 97.00 %, with a recall of 96.64 % and an$F 1$-score of 96.39 %. By comparison, the CBAM model has a reported accuracy of 95.34 %. These results indicate the successful integration of attention modules with GhostNet to support accuracy in diagnostic prediction; SE performs best with ECG image-based heart disease classification and prediction.

Fields of Study & Subjects

ECG Monitoring and Analysis Artificial Intelligence in Healthcare Non-Invasive Vital Sign Monitoring

Publication details

Journal / source
International Conference on Communication, Computing and Electronics Systems (ICCCES)
Pages
1–8
Publication date
21 January 2026
Language
EN