Image Processing Reference
In-Depth Information
CHAPTER 4
Fringe noise removal of
retinal fundus images using
trimming regions
Arezoo Ektesabi; Ajay Kapoor
Swinburne University of Technology, Melbourne, Victoria, Australia
Abstract
Recent advancements in ophthalmology have been significantly influenced by enhancements in image-
processing procedures. Automatic detection of main regions of the eye has been of interest for years,
resulting in multiple studies, methodologies for accurate and fast detection of retinal features. Despite
the improvements in procedures, majority of the reported techniques cannot perform as accurately in
response to the change of the image resolution and size. Therefore, the automated process needs to have
the capability of recognizing the input image and irrespective of its resolution detect the required feature
with high accuracy. To achieve this, it is essential to remove all possible noises such as the effect of the
illuminated light at the edges of the retinal images which in particular affect the results to be inadequate
for localization of the optic disk. This study proposes the use of circular and elliptical trimming regions
for the removal of these bright noises. Significant improvements have been reported in the outcomes for
optic disk localization.
Keywords
Fundus image
Image processing
Noise removal
Ophthalmology
Trimming region
1 Introduction
For centuries with the rapidly growing population, the need for improvements in medical and
health care systems has been felt. As one of the main sensory organs, eye has always played
an important role in any individuals' life. Therefore, ophthalmological research has been of in-
terest and closely monitored areas, especially since the 1999, with the introduction of the vision
Search WWH ::
Custom Search