Date Approved
3-8-2023
Embargo Period
3-8-2023
Document Type
Thesis
Degree Name
M.S. Electrical and Computer Engineering
Department
Electrical and Computer Engineering
College
Henry M. Rowan College of Engineering
Sponsor
National Science Foundation
Advisor
Nidhal C. Bouaynaya, Ph.D.
Committee Member 1
Ghulam Rasool, Ph.D.
Committee Member 2
Ravi Ramachandran, Ph.D.
Keywords
Artificial intelligence, Computer vision, Machine learning, Medical AI, Transformers
Subject(s)
Computer vision in medicine; Diagnostic imaging
Disciplines
Bioimaging and Biomedical Optics | Electrical and Computer Engineering
Abstract
The field of medical imaging has seen significant advancements through the use of artificial intelligence (AI) techniques. The success of deep learning models in this area has led to the need for further research. This study aims to explore the use of various deep learning algorithms and emerging modeling techniques to improve training paradigms in medical imaging. Convolutional neural networks (CNNs) are the go-to architecture for computer vision problems, but they have limitations in mapping long-term dependencies within images. To address these limitations, the study explores the use of techniques such as global average pooling and self-attention mechanisms. Additionally, the study investigates the performance of vision transformers (ViTs), which have shown potential for outperforming CNNs in image classification tasks. The Scopeformer, a new end-to-end architecture that combines the unique strengths of both CNNs and ViTs, is proposed to improve upon their individual performance. The study contributes to the conversation about effective approaches for tackling challenging computer vision tasks in medical imaging.
Recommended Citation
Barhoumi, Yassine, "Efficient Scopeformer: Towards Scalable and Rich Feature Extraction for Intracranial Hemorrhage Detection using Hybrid Convolution and Vision Transformer Networks" (2023). Theses and Dissertations. 3086.
https://rdw.rowan.edu/etd/3086
