Greyscale analysis of retina scans

Che Azemin, M 2012, Greyscale analysis of retina scans, Doctor of Philosophy (PhD), Electrical and Computer Engineering, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Che_Azemin.pdf Thesis Click to show the corresponding preview/stream application/pdf;... 2.81MB
Title Greyscale analysis of retina scans
Author(s) Che Azemin, M
Year 2012
Abstract This study explores the analysis of retina images at the greyscale level, in particular the discovery of a novel clinical marker to describe retinal complexity based on the Fourier fractal dimension (FFD). This work demonstrates that using the box-counting technique on vessel segmentation to summarize retinal complexity can be unreliable, due to the presence of imaging artefacts such as shadow occlusion and imbalanced illumination. The newly developed fractal methodology has been applied successfully and found to be associated with different physiological factors and conditions. The FFD feature has also been demonstrated in a population study to be able to predict stroke incidence 10 years before the actual event. This is an important finding because the current technique is only able to describe the difference between stroke subtypes, but gives no evidence in the prospective data. This research work also shows that vessel segmentation causes imprecision in vessel diameter measurements. To compensate for the segmentation error, a new methodology has been proposed at the greyscale level. The validity of the diameter measurement is confirmed by the observation of cardiac trends in the vessel calibre. The results not only show significant diameter change across the cardiac cycle, but also indicate a specific trend for the arteriole and venule.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Electrical and Computer Engineering
Keyword(s) Retina
Clinical Marker
Version Filter Type
Access Statistics: 229 Abstract Views, 244 File Downloads  -  Detailed Statistics
Created: Tue, 18 Jun 2013, 09:42:58 EST by Brett Fenton
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us