Classification of Breast Cancer Histology Using pretrained CNNs

Publication Date

Spring 4-2025

Presentation Length

Poster/Gallery presentation

College

College of Sciences & Mathematics

Department

Math and Computer Science, Department of

Student Level

Undergraduate

SPARK Category

Research

Faculty Advisor

Chistina Davis

SPARK Session

11:45-12:45, 10:15-11:45 Poster/Gallery Session

Presentation Type

Poster

Summary

Breast cancer is one of the most deadly cancers among women, while analysis of biopsies remains the gold standard for diagnosis. Thus, integration of Convolution Neural Networks (CNNs) into the analysis and classification of histology imaging has significantly advanced the field of pathology. This study will focus on developing a CNN to classify histological images to assist in breast cancer diagnoses. More specifically, we will be leveraging transfer learning via specified pre-trained models on a dataset of Hematoxylin and Eosin (H&E) stained breast biopsy microscopy images. This approach utilizes fine-tuning of pretrained models, image processing and augmentation, and dynamic learning rate scheduling.

This document is currently not available here.

Share

COinS