IoT-based ECG monitoring for arrhythmia classification using Coyote Grey Wolf optimization-based deep learning CNN classifier. (July 2022)

• Record Type: Journal Article
• Authors: Kumar, Abhishek; Kumar, SwarnAvinash; Dutt, Vishal; Dubey, Ashutosh Kumar; García-Díaz, Vicente
• Is Part Of: Biomedical signal processing and control. Volume 76(2022)
• Publication Date: 2022-07
• Subjects: Electrocardiogram -- IoT -- Hybrid optimization -- Arrhythmia classification -- Deep learning network; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.103638
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: In the smart healthcare applications, the arrhythmia classification in patients is performed using the ECG signals collected through the IoT nodes. The arrhythmia classification is performed using the proposed Coy-Grey Wolf optimization-based deep convolution neural network (Coy-GWO-based Deep CNN) classifier from the ECG features. The proposed Coy-Grey Wolf optimization is developed through hybridizing the social prevalence characteristics and the hierarchy-based hunting characteristics of the Canidae family. The proposed Coy-GWO-based Deep CNN's classification accuracy is 95%, showing an improved performance than the existing techniques. Abstract: An electrocardiogram (ECG) is extensively used to evaluate the heart condition that can lead to further investigate heart ailments detection of heart diseases. The process is simple, quick, and non-invasive. However, manually detecting and classifying arrhythmia is not easy, as manually analyzing the ECG signals is time-consuming. This research proposes an automatic detection and classification of arrhythmia using the proposed optimized deep learning classifier. Initially, the ECG signals collected using the Internet of Things (IoT) nodes are processed to generate the QRS complex and the RR interval for establishing the feature vector for the arrhythmia classification, which is done using the proposed Coy-Grey Wolf optimization-based deep convolution neural network (Coy-GWO-based Deep CNN) classifier that detects the anomalies in the ECG signal. The proposed Coy-GWO algorithm inherits the hybrid characteristics, such as social hunting hierarchies and the hunting experiences of the canids, which ensures the effective parameter update in the classifier. Finally, the classification model is implemented and analyzed based on the performance measures. The proposed Coy-GWO-based Deep CNN attained the classification accuracy of 95%, which outperforms the existing techniques.