ORCID

https://orcid.org/0009-0008-4769-2370

Year

2023

Season

Spring

Paper Type

Master's Thesis

College

College of Computing, Engineering & Construction

Degree Name

Master of Science in Computer and Information Sciences (MS)

Department

Computing

NACO controlled Corporate Body

University of North Florida. School of Computing

First Advisor

Dr. Xudong Liu

Second Advisor

Dr. Indika Kahanda

Third Advisor

Dr. Robert Maxwell

Abstract

Breast density screenings are an accepted means to determine a patient's predisposed risk of breast cancer development. Although the direct correlation is not fully understood, breast cancer risk increases with higher levels of mammographic breast density. Radiologists visually assess a patient's breast density using mammogram images and assign a density score based on four breast density categories outlined by the Breast Imaging and Reporting Data Systems (BI-RADS). There have been efforts to develop automated tools that assist radiologists with increasing workloads and to help reduce the intra- and inter-rater variability between radiologists. In this thesis, I explored two deep-learning-based approaches on breast density classification. First, I developed and experimented with algorithms using deep learning models (such as Inception V3 and ViT) to classify patients according to BI-RADS using various types of digital mammograms. Second, with the need to provide not only such classification but also a quantitative measure of breast density to help standardize assessments across radiologists, I applied a deep learning based semantic segmentation model, DeepLabV3, to predict density percentages which then were used to provide a linear and probability scale.

Download

Share

COinS