Comparative Study of methodologies used to identify microaneurysms in RETINA

Image Processing

Diabetes mellitus, commonly known as diabetes has become a huge problem in recent days throughout the world. If the patient’s sugar levels in blood are not maintained properly, it may lead to deformities in the retina over a long period of time. These deformities are diagnosed as Diabetic Retinopathy (DR). Some major problems associated with it are Hard Exudates, Cottonwool spots, hemorrhages & Microaneurysms. The topic of study relates to early detection of such microaneurysms in the retina of the patient diagnosed with Diabetic Retinopathy. Microaneurysms are capillary dilatations that are circular and appear as red dots in the retina. If not treated within time, these lesions may result in loss of vision since it is directly associated with the macula of the eye. These lesions are generally present in inner nuclear layer i.e., the 4th layer of retina. Many methodologies have been proposed for detection of microaneurysms. However, each of them is lacking one thing or the other such as inadequate database, poor pre-processing, false identification etc., In this study, a comparative study between different methodologies that have been implemented and various characteristics that are associated with them is carried out. Comparison of these methods along with evaluating their performance would help in identifying the problems faced by each of these methods. That would give an insight on how to proceed with if the implementation is picked up after this comparative study. Relative differences present in them can be sorted out and used for the advantage of the method that is selected from the pool of methodologies available.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.