Date of Award

Spring 2024

Document Type

Thesis - SCU Access Only

Publisher

Santa Clara : Santa Clara University, 2023

Department

Bioengineering

First Advisor

Yuling Yan

Second Advisor

Hamed Akbari

Abstract

Cardiomyopathy refers to conditions that make it harder for the heart muscles to contract, which may manifest as muscle thickening, stiffening, or growing abnormally. COVID-19 has been associated with cardiovascular complications like cardiomyopathy and heart failure. Several studies have found that 20-30% of COVID-19 patients may develop cardiomyopathy or experience heart failure, especially in severe cases. Detection of cardiac complications such as cardiomyopathy is vital because it can be fatal. If caught in the early stages, it can be mitigated with noninvasive strategies like medications and lifestyle changes, rather than surgical implants and potentially transplants. Unfortunately, there can be barriers to accessing a diagnosis in a timely and cost-effective manner. Additionally, low income, whether individual or regional, can heavily impact the ability to get diagnostic tests.

Machine learning can offer earlier detection, as it has been used to detect hypertrophic cardiomyopathy in multiple age groups with high specificity and sensitivity. For our project, we will use machine learning to detect cardiomyopathy in COVID-19 patients. We will use data that involves COVID-19 patients who have cardiomyopathy and COVID-19 patients without cardiomyopathy to train a model for detection. Our main objectives are to preprocess the ECG image data to remove extraneous information for improved model performance and use the transfer learning method and a deep learning model to classify cardiomyopathy in COVID-19 patients. Our project is necessary because it will potentially help save lives and can be used by many different types of people such as healthcare professionals.


Download

SCU Access Only

To access this paper, please log into or create an account in Scholar Commons using your scu.edu email address.

Share

COinS