Mohammad, Fakher (2022) Machine Learning Models for Recognizing Curve-shaped Abnormalities in Different Image Modalities. Doctoral thesis, The University Of Buckingham.
![]() |
Text
Mohammad_Final_Thesis_148976_955198314.pdf - Submitted Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (15MB) |
Abstract
This thesis is concerned with using machine learning algorithms for the analysis of different image modalities for the presence of abnormal features or shapes. This challenge appears in various crucial computer vision applications in science, engineering, medicine, and art. Different image modalities abnormal features/regions are often very specific to the applications, the capturing tools, and the subjects of the scenes that were captured. The appearance of certain types of feature abnormalities often indicates potentially serious faulty defects in the imaged objects. We only consider two applications: Inspecting cracks in building materials such as glass façades and concrete surfaces using digital camera images and determining irregularity properties of tumour lesion borders from ultrasound (US) scan images. In the first case, abnormalities appear as cracks that could endanger life and infrastructure. At the same time, irregularity of the tumour border, which is highly correlated to the malignancy of the tumour, could reduce the patient's recovery chances if not recognised/treated early...
Item Type: | Thesis (Doctoral) |
---|---|
Uncontrolled Keywords: | Machine learning ; image modalities ; image abnormalities ; cancer ; CNN models ; AI methods. |
Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software R Medicine > R Medicine (General) T Technology > T Technology (General) |
Divisions: | School of Computing |
Depositing User: | Freya Tyrrell |
Date Deposited: | 19 May 2025 08:14 |
Last Modified: | 19 May 2025 08:14 |
URI: | http://bear.buckingham.ac.uk/id/eprint/667 |
Actions (login required)
![]() |
View Item |