The University of Edinburgh
Development and clinical translation of optical and software methods for endomicroscopic imaging
Abstract
dc:description.abstractEndomicroscopy is an emerging technology that aims to improve clinical diagnostics by allowing for in vivo microscopy in difficult to reach areas of the body. This is most commonly achieved by using coherent fibre bundles to relay light for illumination and imaging to and from the area under investigation. Endomicroscopy’s attraction for researchers and clinicians is two-fold: on the one hand, its use can reduce the invasiveness of a diagnostic procedure by removing the need for biopsies; On the other hand, it allows for structural and functional in vivo imaging. Endomicroscopic images acquired through optical fibre bundles exhibit artefacts that deteriorate image quality and contrast. This thesis aims to improve an existing endomicroscopy imaging system by exploring two methods that mitigate these artefacts. The first, software-based method takes several processing steps from literature and implements them in an existing endomicroscopy device with a focus on real-time application to enable clinical use, after image quality was found to be inadequate without further processing. A contribution to the field is that two different approaches are implemented and compared in quantitative and qualitative means that have not been compared directly in this manner before. This first attempt at improving endomicroscopy image quality relies solely on digital image processing methods and is developed with a strong focus on real-time applicability in clinical use. Both approaches are compared on pre-clinical and clinical human imaging data. The second method targets the effect of inter-core coupling, which reduces contrast in fibre images. A parallelised confocal imaging method is developed in which a sequence of images is acquired while selectively illuminating groups of fibre cores through the use of a spatial light modulator. A bespoke algorithm creates a composite image in a final processing step. In doing so, unwanted light is detected and removed from the final image. This method is shown to reduce the negative impact of inter-core coupling on image contrast on small imaging targets, while no benefit was found in large, scattering samples.
Degree
thesis:*- Grantor dc:publisher
- The University of Edinburgh
- Year dc:date.issued
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Norberg, Dominic R.T.
- Advisors dc:contributor.advisor
-
- Dhaliwal, Kevin
- Krstajic, Nik
Subjects
dc:subject × 3Rights
- Language dc:language.iso
- en