Cognitive assistant DeepNet model for detection of cardiac arrhythmia

Deep Learning (DL) has become a topic of study in various applications, including healthcare. The important factors considered in building a deep learning model are hardware, low power consumption and high accuracy. Workstations with high level graphics processing units (GPU) are commonly used to achieve high performance in deep learning applications. However, it is a lot in terms of cost and power consumption to build a highperformance platform. In this study, a single-board computer- NVIDIA Jetson nano developer kit is used for training and testing the proposed model. A novel deep learning model is proposed that facilitates automated classification of electrocardiogram (ECG) into seven types of ECG beats. A novel structural CNN unit is constructed using pointwise convolutions and depthwise separable convolution. Structural CNN units with different kernels are used to design our proposed DeepNet model. Two pretrained models CNN models, AlexNet, Resnet18 are also explored and results are compared to evaluate efficiency of our model. Models are evaluated on MIT-BIH arrhythmia database. Moreover, our proposed DeepNet model is effective, promising, outperformed all State of Art Models with an accuracy of 99.56%.