A Comprehensive Survey on Heart Disease Detection Using Deep Learning and Benchmark Datasets

Heart diseases are increasing and have become a serious health problem. There are different types of heart diseases, and each type has several factors that affect the heart. For this, early diagnosis is important, but the existing methods and treatments are time-consuming. To overcome challenges like this, some studies propose different types of automated models that can assist doctors in real time. Those studies used machine learning (ML), convolutional neural networks (CNN), and deep learning (DL) architectures. In the process of heart disease detection, playing a significant role to get the best accurate results and the best datasets is also important. Based on the datasets, the model's performance will improve, and also some of the datasets are updated every year; those are also known as benchmark datasets. This survey represents ImageCAS, ACDC, ARCAD, MIMIC, and CCTA are 5 benchmark datasets that are widely used for different tasks and mainly used for heart disease detection using various models. Still, the process is continuing on upgrading the datasets; these are achieving promising results.