Deep Learning Techniques for Breast Ultrasound Image Classification and Segmentation
Ryan Jimenez
Co-Presenters: Individual Presentation
College: The Dorothy and George Hennings College of Science, Mathematics and Technology
Major: Computer Science
Faculty Research Mentor: Meng Xu
Abstract:
Checking for breast cancer is one of the most time-consuming and expensive processes, despite it being one of the leading causes of cancer-related deaths globally. This study explores the application of deep learning techniques within the frameworks of classification and segmentation to improve the efficiency and accuracy of breast cancer detection. We compared multiple approaches for single-task learning models and ultimately ended with the approach of a multitasking learning model that performed both tasks simultaneously. The results shown are from the models we experimented with; although the results aren’t pixel-perfect, they still highlight the potential of deep learning and its connection to the breast cancer diagnostic process. If successfully implemented, it has the potential to decrease the time and cost it takes to perform these tests significantly, prompting individuals to get checked more often and potentially leading to earlier detection and treatment.