Key result
Convolutional Neural Networks (CNN) were identified as the most suitable and frequently used deep learning method for feature extraction and arrhythmia classification.
Why the study?
Cardiovascular disease mortality is projected to peak globally in India by 2030, highlighting the need for timely detection and accurate deep learning models for arrhythmia classification.
Convolutional Neural Networks (CNN) are highly effective for ECG feature extraction and arrhythmia classification, achieving high accuracy in detecting conditions like atrial fibrillation.
Answers are grounded in this paper and cite their sources.
Supports CNN prioritization for ECG arrhythmia algorithms; leaves open prospective clinical validation before adoption.
National Conference for doctors of HAL on May 2022 published in its report that, India will record the maximum number of deaths due to cardiovascular diseases (CVDs) by 2030 in the world. So its timely detection and cure can save many lives. To detect and classify CVDs, deep learning (DL) methods are widely used. This survey focuses on a variety of DL models applied in various research papers, to find models with higher accuracy in the classification of arrhythmia and other heart-related issues. Here six DL models have been focused namely Convolution Neural Network (CNN), Gated Recurrent Unit (GRU), Long Short Term Memory (LSTM), Multi-Layer Perceptron (MLP), Deep Belief Network (DBN), and Recurrent Neural Network (RNN) after the review of 23 papers from the year 2018-2022, to find the suitable model for classification task and feature extraction where Electrocardiogram (ECG) is a common input to the every DL model.
No takes yet. Share an insight, caveat, or question.
Sharma et al. (2023) conducted a review in Arrhythmia. Deep Learning Methods (CNN, GRU, LSTM, MLP, DBN, RNN) vs. Traditional methods or other deep learning models was evaluated on Classification accuracy. Convolutional Neural Networks (CNN) were identified as the most suitable and frequently used deep learning method for feature extraction and arrhythmia classification.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: