Manual Tool and Semi-Automated Graph Theory Method for Layer Segmentation in Optical Coherence Tomography

Sayers, Dean and Habib, Maged and Al-Diri, Bashir (2019) Manual Tool and Semi-Automated Graph Theory Method for Layer Segmentation in Optical Coherence Tomography. In: Science and Information (SAI) Conference, 16-17 July 2019, London, United Kingdom.

Documents
Manual Tool and Semi-Automated Graph Theory Method for Layer Segmentation in Optical Coherence Tomography

Request a copy
[img] PDF
SAI Conference final.pdf
Restricted to Repository staff only

1MB
Item Type:Conference or Workshop contribution (Paper)
Item Status:Live Archive

Abstract

Optical Coherence Tomography (OCT) is a major tool in the diagnosis of various diseases. Disease diagnosis is based on various features within the OCT images, including retinal layer positions and the distances between them and the build-up of fluid. All of these features require an expert marker in order to identify them so that the information can properly aid in the diagnosis for the patient. This process takes an incredible amount of time for the expert carry out as they need to manually trace the layers for every frame. This therefore indicates that there is a need for automation so that the expert can more easily and efficiently label the retinal layers. In this project two processes were developed. The first step is to use a semi-automated graph theory method to segment a specific layer given a rectangular region of interest, specified by the user. The output of the first process can then be corrected, where needed, using the manual tool. This method can segment layers with on average less than 1-2 pixels of error vs two expert markers.

Keywords:OCT Image, Graph Theory, Semi Auto Manual
Subjects:G Mathematical and Computer Sciences > G400 Computer Science
Divisions:College of Science > School of Computer Science
ID Code:35509
Deposited On:09 Apr 2019 10:08

Repository Staff Only: item control page