Advancing Biomedical Image Segmentation of Lower-Grade Gliomas using Transfer Learning

Vanapalli, Yaswanth (2024) Advancing Biomedical Image Segmentation of Lower-Grade Gliomas using Transfer Learning. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (911kB) | Preview

Abstract

This study aims to enhance the segmentation of LGGs in MRI scans using transfer learning, particularly transformer-based pretrained models. Medical imaging is challenging with LGGs due to their complex and diffuse nature, as they are classified as brain tumors. There is an issue with traditional approaches of how these tumors are segmented which tends to be so much time consuming and very subjective. To address these challenges, this study leverages the Swin Transformer, a robust vision transformer, to enhance segmentation efficiency and precision without demanding significant annotated datasets and computational power. The proposed method adapts the Swin Transformer model that is trained for large image data, to detect tumor regions accurately. The model was assessed based on four main measures such as the Dice Similarity Coefficient (DSC), Intersection over Union (IoU), and classification accuracy of a test set. The experimental evaluation exhibits high efficiency with a test accuracy of 99.64%, the mean IoU of 0.837, and the mean Dice score of 0.861. However, it was observed that a few imperfections exist in the model, and those are mainly related to the recall of smaller or more intricate tumor regions. This study establishes the viability of transformer based models in medical image segmentation and offers a solid starting point for improving LGG diagnosis in healthcare facilities.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Jilani, Musfira UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > Biomedical engineering Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence > Computer vision Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence > Computer vision Q Science > Life sciences > Medical sciences > Pathology > Tumors
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	05 Sep 2025 13:32
Last Modified:	05 Sep 2025 13:32
URI:	https://norma.ncirl.ie/id/eprint/8830

Actions (login required)

View Item