Retinal Image Processing

We are attempting to characterise and quantify the effects of various diseases on the retina. Primarily we are interested in glaucoma and diabetes, as they give rise to the most easily observed changes and are significant causes of loss of vision.

The images are characterised by:


Glaucoma is the major cause of loss of sight in the registered blind population. It is characterised by raised intra-ocular pressure (IOP) caused by blockages in structures around the lens preventing vitreous humour leaking out as it should. Without treatment nerve damage occurs at the optic nerve head which has the effect of creating areas of low vision. These are randomly shaped and randomly positioned over the retina and are often not noticed by the patient until the vision loss is quite severe, and irreversible.

Diagnosis of glaucoma is by

Once diagnosed, glaucoma can be treated with eye drops, or surgery if these are ineffective. Either way, frequent and regular monitoring is required.

This project is concerned with producing software tools to aid in the automated diagnosis and monitoring of the disease. These are tasks which are presently done by an expert viewing images of the optic nerve head, such as these healthy and diseased eyes. (The images are negatives.)

Several years ago, a student assessed the consistency of experts in drawing the optic nerve boundary. He sent ten experts ten images each to analyse, they returned ten slightly different results. His conclusion was that accuracy is impossible because we all define the boundary differently, but consistency is paramount.

normal eye

Click an image to enlarge eye with glaucoma

The images show the region of the retina where nerve fibres and blood vessels pass through the eye, i.e. the blind spot. In a normal, healthy eye, the nerve head region is approximately circular. In an eye effected by glaucoma, nerve damage results in a change to a more irregular shape.

Identifying the structures is difficult as they are indistinct, of lower contrast than and partially obscured by blood vessels.

Our initial work resulted in a semi automatic method of identifying the nerve head boundaries. It worked by asking the user to indicate the approximate location of the boundary in a median filtered version of the image. The software moves the user's points to a consistent location on the boundary. It also inserts additional points by interpolation. The boundary between the points is tracked. The work was presented at the American Academy of Optometrists Biennial Meeting in Europe in 1994 and at the 2nd Internet World Congress on Biomedical Science.

A major disadvantage of this approach is its sensitivity to the user's initial selection of points. The intra observer variability in the latest version is in the order of 4 % (in a normal eye, the variability is typically 8% and in a diseased eye 4%). This may be reduced to zero by automating the boundary point idetification.

We have worked on multiresolution methods incorporating the wavelet transform and steerable difference of Gaussian filters to enhance the images.

We have described in a 1999 paper how a dynamic contour (snake) may be shrunk onto the boundaries in the optic nerve head. The dynamic contour requires that the image is enhanced to increase the contrast between the nerve head and the retinal regions. The snake can then shrink reliably onto the boundary.

eye with dynamic contour Click the image to enlarge

We are currently investigating how the texture (if any) in the image is correlated with the anatomy. More examples will follow in due course.


Diabetes affects the circulatory system as vessaels become porous and leak blood and other fluids into retinal tissue. This is characterised by changes to the appearance, where blood or lipids congeal. Unless controlled, the illness can, and does, lead to vision loss. The incidence of heart attack, stroke and reanl disease is also higher than average in diabetics since these are complications caused by by the compromised circulatory system.

There is a school of thought that suggests that there may be a relationship between retinal appearance and the incidence of heart attack and stroke since they may be causally related.

More info to follow in due course.


We are currently investigating BRIEF (Binary Robust Independent Elementary Feature) as a tool for assessing texture.

Results will be posted as they become available.

This work has been partially funded by the International Glaucoma Association.

Last modified
July 2012