HeartNet: Self Multihead Attention Mechanism via Convolutional Network With Adversarial Data Synthesis for ECG-Based Arrhythmia Classification

Abstract

Cardiovascular disease is now one of the leading causes of morbidity and mortality. Electrocardiogram (ECG) is a reliable tool for monitoring the health of the cardiovascular system. Currently, there has been a lot of focus on accurately categorizing heartbeats. There is a high demand for automatic ECG classification systems to assist medical professionals. But there is a big issue in obtaining original data extensively in medical domains in rare diseases, so it is essential to have a robust solution adopting this challenge. So, we need a solution that can address the problem of tackling data insufficiency, which is a major concern nowadays for medical applications. Without having, significant training samples the overall output can be demised. But the recent works on ECG classification did not address the challenge of solving the data insufficiency label problem. To overcome this issue, we developed a new generative adversarial network-based deep learning method called HeartNet for tackling the data insufficiency problem. The proposed deep learning method is compressed by a multi-head attention mechanism on CNN architecture. The main challenge of insufficient data labels is solved by adversarial data synthesis by adopting a generative adversarial network (GAN) with generating additional training samples. It drastically improves the overall performance of the proposed method by 5-10% on each insufficient data label category. Since the training samples are increased. We evaluated our proposed method utilizing the MIT-BIH dataset. Our proposed method has shown $99.67\pm {0.11}$ accuracy and 89.24± 1.71 MCC trained with adversarial data synthesized dataset. However, we have also utilized two individual datasets as Atrial Fibrillation Detection Database and PTB Diagnostic Database to see the performance and generalization of our proposed model on ECG classification. The effectiveness and robustness of the proposed method are validated by extensive experiments, comparison, and analysis. Later on, we also highlighted some limitations of this work.