Electrocardiogram heartbeat classification based on a deep convolutional neural network and focal loss

• Published in: Computers in biology and medicine Vol. 123; p. 103866
• Main Authors: Romdhane, Taissir Fekih, Alhichri, Haikel, Ouni, Ridha, Atri, Mohamed
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.08.2020; Elsevier Limited
• Subjects: AAMI standard; Accuracy; Algorithms; Arrhythmia; Artificial neural networks; Back propagation; Cardiac arrhythmia; Cardiovascular diseases; Classification; Convolutional neural network; Datasets; ECG Classification; Echocardiography; EKG; Electrocardiography; Feature extraction; Focal loss; Heart; Heartbeat category; Imbalanced data; INCART; Internal Medicine; Machine learning; Methods; MIT-BIH; Morphology; Neural networks; Optimization; Other; Segmentation; Signal processing; Support vector machines; Wavelet transforms; Imbalanced data; ECG Classification; AAMI standard; Convolutional neural network; Heartbeat category; INCART; MIT-BIH; Focal loss
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2020.103866

The electrocardiogram (ECG) is an effective tool for cardiovascular disease diagnosis and arrhythmia detection. Most methods proposed in the literature include the following steps: 1) denoizing, 2) segmentation into heartbeats, 3) feature extraction, and 4) classification. In this paper, we present a deep learning method based on a convolutional neural network (CNN) model. CNN models can perform feature extraction automatically and jointly with the classification step. In other words, our proposed method does not require a feature extraction step with hand-crafted techniques. Our proposed method is also based on an algorithm for heartbeat segmentation that is different from most existing methods. In particular, the segmentation algorithm defines each ECG heartbeat to start at an R-peak and end after 1.2 times the median RR time interval in a 10-s window. This method is simple and effective, as it does not use any form of filtering or processing that requires assumptions about the signal morphology or spectrum. Although enhanced ECG heartbeat classification algorithms have been proposed in the literature, they failed to achieve high performance in detecting some heartbeat categories, especially for imbalanced datasets. To overcome this challenge, we propose an optimization step for the deep CNN model using a novel loss function called focal loss. This function focuses on minority heartbeat classes by increasing their importance. We trained and evaluated our proposed model with the MIT-BIH and INCART datasets to identify five arrhythmia categories (N, S, V, Q, and F) based on the Association for Advancement of Medical Instrumentation (AAMI) standard. The evaluation results revealed that the focal loss function improved the classification accuracy for the minority classes as well as the overall metrics. Our proposed method achieved 98.41% overall accuracy, 98.38% overall F1-score, 98.37% overall precision, and 98.41% overall recall. In addition, our method achieved better performance than that of existing state-of-the-art methods. Graphical Abstracts should summarize the contents of the article in a concise, pictorial form designed to capture the attention of a wide readership online. Refer to the following website for more information: http://www.elsevier.com/graphicalabstracts Thank you for your interest in our journal, and best wishes for good progress in your research endeavours. Include interactive data visualizations in your publication and let your readers interact and engage more closely with your research. Follow the instructions here: https://www.elsevier.com/authors/author-services/data-visualization to find out about available data visualization options and how to include them with your article. [Display omitted] •A new method for segmenting ECG signals that does not require any denoising, or resampling of ECG signals.•Using the latest advanced design and training techniques to build a deep CNN model for classification.•Exploiting a novel type of loss function called focal loss to handle the imbalanced classes problem.•Using MIT-BIH, the method achieved 98.41%, 98.86, and 98.38% for the average accuracy, AUC, and F1-score respectively.