Key result
C-BiNET model achieves ~98% accuracy in categorizing five arrhythmia types.
Why the study?
The delicate patterns and non-linear nature of ECG signals make rapid, accurate manual classification challenging, creating a need for automated systems to detect abnormal heart rhythms.
Does the C-BiNET deep learning model improve the accuracy of arrhythmia classification from ECG signals compared to standard CNN?
Does the C-BiNET deep learning model improve the accuracy of arrhythmia classification from ECG signals compared to standard CNN?
The C-BiNET deep learning model demonstrates high accuracy in automating the classification of arrhythmias from ECG signals, highlighting its potential utility for real-time clinical monitoring.
Arrhythmia, an irregular cadence of the heartbeat, with its frequency growing as individuals age. An electrocardiogram (ECG) serves as a usual tool for monitoring cardiac mechanics. The delicate patterns, and non-linear nature of ECG signals make rapid and accurate manual classification highly challenging. To address this, an automated system capable of perceiving various abnormal heart tempos within extensive ECG data sets is essential for healthcare management system. This study presents the C-BiNET model, a deep learning strategy that integrates convolutional neural networks (CNNs) and bidirectional long short-term memory (Bi-LSTM) networks to categorize ECG signals into five distinct categories in real time. The C-BiNET model encompassing four key stages: signal preprocessing, data augmentation, feature extraction, and classification. To refine the functionality of our system, the input signals undergo preprocessing using wavelet transform, effectively eliminating noise from the source data. The extraction of R-peak information following denoising facilitates the individual heartbeat separation. Subsequent application of the ADASYN algorithm augments the dataset, bolstering the model's robustness. Then CNN and Bi-LSTM is utilized for feature extraction to capture discriminative patterns. The final classification is accomplished through fully connected (FC) layers with a softmax function. This model successfully categorizes five types of arrhythmias utilizing the MIT-BIH arrhythmia database (MITArrDB), achieving an impressive F1-score of 98.35% and a classification accuracy of 98.24%. Comparative analysis with standard CNN exposes the superior performance of proposed model and highlights potential utility in clinical applications.
No takes yet. Share an insight, caveat, or question.
Febeena et al. (2025) studied Arrhythmia. C-BiNET model (CNNs and Bi-LSTM) vs. Standard CNN was evaluated on Classification accuracy and F1-score for categorizing ECG signals into five distinct arrhythmia categories. The C-BiNET model successfully categorized five types of arrhythmias using the MIT-BIH database, achieving an F1-score of 98.35% and a classification accuracy of 98.24%.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: