Early Detection of Aortic Stenosis for the Prevention of Heart Failure in High-Risk Population using Whale Optimization Algorithm

Abstract

Currently, the pace of life of people in a Smart City has been incrementally affected due to several factors such as stress and risk diseases, including hypertension, diabetes, and even SARS. COVID'19 pandemic with a future prognostic of 7,87 million deaths to the middle of 2022- that leads to cardiovascular diseases and conduct to a heart attack, it has been observed that in Mexico it is recurrent from the range age near of 45 and one of the most common heart diseases is aortic stenosis and can reach be fatal if it is not detected in time. In practice, the echocardiogram is the fastest means to detect abnormalities anatomically and physiologically in real-time. However, the interpretation can be affected by the image quality. Therefore, there are techniques to improve noise for image processing and, where appropriate, separate the regions of interest. This is known as segmentation. In this research, a model inspired by artificial intelligence for diagnosis is proposed, the model uses Convolutional Neural Networks (CNN) as a deep learning technique for classifying images of Aortic Stenosis, and an innovative metaheuristic named Whale Optimization Algorithm (WOA) is implemented to select the most relevant segmentation feature of the Aortic echocardiographic view. The results show promising performance for the diagnosis.