How to find lesions in a retinal image having no specific shape and colour

I am working with retinal images from where I need to detect the soft exudates (cotton wool spots) but the problem is that the cotton wool spots doesn't have any specific shape, colour or location. It may occur in any portion of the retinal image. Its shape also varies in every image. I tried with the thresholding technique but it failed. Along with the cotton wool spots, other features like optic disk and hard exudates are also detected. This is because its intensity is similar to optic disk and hard exudates. Can anyone suggest me any idea to detect only the cotton wool spots?

Answer 1

[Image Analyst]

Look under the eye section here: http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models I'm sure you'll find papers with algorithms there. Pick one and code it up. There is no exudates finding function in MATLAB so you have no choice but to build it up from lower level functions.