Towards Automatic Analysis on Refraction Related Tissue of the Anterior Segment

Zhang, Yang

Towards Automatic Analysis on Refraction Related Tissue of the Anterior Segment

Zhang, Yang

Permalink

Publication Type:: Thesis
Issue Date:: 2022

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download contents and abstractAdobe PDF (175.53 kB)

Adobe PDF

Download thesisAdobe PDF (12.66 MB)

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Zhang, Yang
dc.date.accessioned	2023-07-11T04:25:17Z
dc.date.available	2023-07-11T04:25:17Z
dc.date.issued	2022
dc.identifier.uri	http://hdl.handle.net/10453/171443
dc.description	University of Technology Sydney. Faculty of Engineering and Information Technology.	en_US.UTF-8
dc.description.abstract	In clinical ophthalmology practice, the tissues derived from the anterior segment play a key role in the light refraction into the eye. Associated analysis for them could provide significant information for the specific disease diagnosis for the clinical experts. While traditional methods tend to investigate these problems with manual operations, of which performances are unsatisfactory. An acceptable solution for that is to exploit the deep learning algorithms, which have obtained tremendous achievements in several medical research areas. This thesis is systematically devoted to the automatic analysis of the refraction associated with the related tissues of the anterior segment. The first task starts with predicting the juvenile refraction power with a designed attentive neural network. Existing standard refraction power acquirement methods based on cycloplegic autorefraction need to induce with specific medicine lotions, which may cause side-effects and sequelae for juvenile students. Besides, several fundus lesions and ocular disorders will degenerate the performance of the objective measurement due to equipment limitations. This thesis tries to deal with this problem by sending ensembled high-level features into an attentive iteration neural network to enhance the representation through an up-bottom fusion path. Then, we turn our glare to the corneal endothelium health status evaluation, which has a great impact on visual acuity and corneal refractivity. The second task is to address the interactive corneal endothelium image segmentation. In clinic medical practice, the well-delineated corneal endothelium images can provide crucial information for evaluating the morphometric parameters. However, the manual annotation of the ground truth is laborious and highly depends on clinical experience, especially when an unseen image modality is exhibited and adequate marked images are not available. This thesis has proposed a weakly supervised siamese network to realize the interactive segmentation, i.e. new modality segmentation, of the endothelium images by exploiting the representative vectors of the target objects.	en_US.UTF-8
dc.format	Thesis (PhD)
dc.language.iso	en_US	en_US.UTF-8
dc.relation	https://opus.lib.uts.edu.au/bitstream/10453/171443/2/02whole.pdf
dc.rights	info:eu-repo/semantics/openAccess
dc.rights	The author owns the copyright in this thesis including all reproduction and reuse rights for the work. The work may not be altered without the permission of the copyright owner. Attribution is essential when quoting or paraphrasing from this thesis.
dc.rights	© 2022 Yang Zhang
dc.rights	au.edu.uts.lib/cph
dc.title	Towards Automatic Analysis on Refraction Related Tissue of the Anterior Segment	en_US.UTF-8
dc.type	Thesis
utslib.copyright.status	open_access	*

Abstract:

In clinical ophthalmology practice, the tissues derived from the anterior segment play a key role in the light refraction into the eye. Associated analysis for them could provide significant information for the specific disease diagnosis for the clinical experts. While traditional methods tend to investigate these problems with manual operations, of which performances are unsatisfactory. An acceptable solution for that is to exploit the deep learning algorithms, which have obtained tremendous achievements in several medical research areas. This thesis is systematically devoted to the automatic analysis of the refraction associated with the related tissues of the anterior segment. The first task starts with predicting the juvenile refraction power with a designed attentive neural network. Existing standard refraction power acquirement methods based on cycloplegic autorefraction need to induce with specific medicine lotions, which may cause side-effects and sequelae for juvenile students. Besides, several fundus lesions and ocular disorders will degenerate the performance of the objective measurement due to equipment limitations. This thesis tries to deal with this problem by sending ensembled high-level features into an attentive iteration neural network to enhance the representation through an up-bottom fusion path. Then, we turn our glare to the corneal endothelium health status evaluation, which has a great impact on visual acuity and corneal refractivity. The second task is to address the interactive corneal endothelium image segmentation. In clinic medical practice, the well-delineated corneal endothelium images can provide crucial information for evaluating the morphometric parameters. However, the manual annotation of the ground truth is laborious and highly depends on clinical experience, especially when an unseen image modality is exhibited and adequate marked images are not available. This thesis has proposed a weakly supervised siamese network to realize the interactive segmentation, i.e. new modality segmentation, of the endothelium images by exploiting the representative vectors of the target objects.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/171443