Author: Noura Hamad Alhawiti
Alhawiti, Noura Hamad, 2018 Bilateral Active Contours, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Active contours (deformable curves), also known as "snakes", were introduced as a general tool for delineating boundaries in computational image analysis. Snakes are considered one of the most successful variational models used in image segmentation. The aim of this method is to segment an image into a finite number of important regions. The theoretical basis of the method is drawn from the calculus of variations theory. Many adaptations of the original method have been introduced over the years using mathematical properties and efficient numerical schemes, usually at the implementation level to improve performance on particular types of problems. Snakes are energy minimising and they balance internal forces that oppose deformation and image forces that pull it towards object contours. This thesis introduces snakes specialised to incorporate bilateral symmetry. The purpose of the study is to develop a theoretical basis for the method of active contours specific to the biological and medical application that directly takes into account first order bilateral symmetry. Human-made objects are often constructed with exact bilateral symmetry; however, in biology, bilateral symmetry is common as a first approximation but not necessarily at all levels of detail. Most vertebrates exhibit bilateral symmetry, as do, for example, the leaves of many plants. This theoretical study proceeds by constructing definitions of functionals and deriving properties analytically. First, an analytical solution are derived from first principles. Second, a discrete solution is derived from the analytical solution to implement the method on real data. Next, a computer program is written in Matlab to implement the bilateral snake. Finally, the computer code is validated on an example image of a human face.
Keywords: Active contours, Geometric active contour, parametric active contour, Kass's Snake, Bilateral active contours.
Subject: Mathematics thesis
Thesis type: Masters
Completed: 2018
School: College of Science and Engineering
Supervisor: Murk Bottema