An Effective Identification of Microaneurysm in Retinal Eye Using ConvNet with Adam Optimizer

This research work outlines a framework to diagnose the Diabetic Retinopathy (DR) from colour fundus images by applying deep learning techniques. ConvNet with Adam Optimization approach is applied for early identification of Microaneurysm occurring in retina of the eye and also for classifying its severity accurately. The data augmentation attribute of ConvNet with Adam Optimiser helps to overcome the complications involved in the categorization task such as non microaneurysm and microaneurysm in human eye. The ConvNet with Adam optimization helps to achieve high accuracy rate, by reducing false positive rate and is demonstrated with the help of high end graphics processor with available kaggle datasets, for a high level classification.