Date Approved

12-12-2024

Graduate Degree Type

Project

Degree Name

Computer Information Systems (M.S.)

Degree Program

School of Computing and Information Systems

First Advisor

Christian Trefftz

Academic Year

2024/2025

Abstract

Alzheimer’s Disease (AD) is a significant and growing global health issue, affecting millions of people around the world. Despite ongoing research, there is currently no cure or highly effective treatment for AD, making early detection essential to slowing its progression and alleviating its impact on patients and healthcare systems. This project explores the use of the "Augmented Alzheimer MRI Dataset V2," a specialized dataset developed to improve the precision and reliability of machine learning models for diagnosing and classifying different stages of AD from MRI scans. The dataset includes MRI images categorized into four classes representing different stages of cognitive impairment: Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented. Traditional datasets often face challenges like limited sample sizes and low variation, which can restrict model effectiveness. To address these issues, this dataset includes both original and augmented images. By applying data augmentation techniques such as rotation, scaling, and flipping the dataset generates diverse image variations while retaining clinically relevant features. This augmented data enhances model training, helping models generalize better and avoid overfitting, which in turn increases resilience to variations in real-world data. The dataset is organized with clear folder structures for both original and augmented images, simplifying the separation of training, validation, and test sets. This structure aids in systematic model development and enables models to learn from a broad range of data types, thereby enhancing their ability to accurately classify stages of AD and advancing the effectiveness of AI in AD diagnosis. In developing this dataset, this project highlights the crucial role of data augmentation in medical imaging to bolster model performance. The "Augmented Alzheimer MRI Dataset V2" marks progress toward creating reliable, AI-based tools for early and accurate detection of Alzheimer’s, aiming to reduce the disease's burden on patients, their families, and healthcare providers worldwide.

Download

Share

COinS