A deep learning-based electrocardiogram classification framework achieved an average accuracy of 97.1%, sensitivity of 97.9%, and specificity of 97.6% for three-class ECG signal classification.
Does a deep learning-based framework improve electrocardiogram signal classification performance under stress conditions?
A deep learning framework using time-frequency signal analysis and transfer learning achieved high accuracy (97.1%) in classifying stress-test electrocardiogram signals.
Accurate interpretation of TM stress-test electrocardiogram signals is important for early identification of cardiovascular abnormalities and continuous healthcare monitoring. This study proposes a deep learning-based electrocardiogram classification framework using time-frequency signal analysis and transfer learning techniques for intelligent healthcare monitoring applications. Electrocardiogram signals obtained from publicly available physiological databases and TM stress-test recordings were transformed into time-frequency representations using continuous wavelet analysis. The generated signal representations were classified using a pre-trained deep learning architecture to improve electrocardiogram signal classification performance under stress conditions. Experimental results demonstrated an average classification accuracy of 97.1 %, sensitivity of 97.9 %, and specificity of 97.6 % for three-class electrocardiogram signal classification. The findings indicate that the proposed framework can support automated electrocardiogram monitoring and intelligent healthcare applications. Nevertheless, the study is limited by dataset size and the absence of external validation. Future work will focus on multicenter electrocardiogram datasets, lightweight deep learning architectures, and intelligent healthcare communication systems integrating software defined radio and visible light communication technologies.
Veerappan et al. (Thu,) conducted a other in Cardiovascular abnormalities. Deep learning-based electrocardiogram classification framework was evaluated on Three-class electrocardiogram signal classification accuracy. A deep learning-based electrocardiogram classification framework achieved an average accuracy of 97.1%, sensitivity of 97.9%, and specificity of 97.6% for three-class ECG signal classification.