Retinal vessel segmentation using Multiheaded attention UNet

Retina is a neural layer consisting of photo-resistors. One of the reasons for retinal damage is Diabetic retinopathy, caused by high blood sugar due to diabetes. If a person is diagnosed with diabetes, it's very important to get regular eye check-ups. Early detection of diabetic retinopathy and early treatment of it would help in stopping the damage and also avert blindness. In order to diagnose it in an early stage, deep learning can play a crucial role. The fundoscopic images of the retina are passed into different stages of the deep-learning model to acquire the twofold output of the retina. The Deep-learning model for this work used is a Deep Convolutional Neural Network. Theimage segmentation can finally interpret whether a person is suffering from diabetic retinopathy and also in which stage itis. This method of deep learning also helps ophthalmologists to verify the results more accurately.