Performing the classification of pulsation cardiac beats automatically by using CNN with various dimensions of kernels

The clinical indication of arrhythmia identifies specific aberrant circumstances in heart pumping that may be detected using electrical impulses during conduction or by allowing a little amount of current to travel through the electrodes, disrupting the cardiac muscle's resistance. The electrocardiogram (ECG) is one of the most important instruments for detecting cardiac arrhythmia since it is the most least intrusive and effective procedure. Physically or visually inspecting the heart is time-consuming and difficult, hence the development of computer aided diagnosis (CAD) is being developed to aid clinical decision-making. In this suggested research, a convolutional neural network (CNN)-based approach is used to automate the heartbeat classification process in order to identify cardiac arrhythmia. The improved enhancement of CNN structure has been implemented in this suggested research. The feature maps are then subjected to the max pooling process. Finally, feature maps are generated by concatenating kernels of different sizes and delivering them as an input to the fully linked layers. The MIT BIH arrhythmia database is used to implement this approach, and the total average accuracy is 99.21%. The proof of the suggested study's efficiency and efficacy in identifying cardiac arrhythmia has also been done via an experimental comparison