Optimised Back Propagation-Based Deep Residual Learning Network Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Images

Diabetic retinopathy is a medical illness characterised by the pathological changes that occur in the retina of individuals diagnosed with diabetes mellitus. This condition is prevalent among a substantial proportion of individuals with diabetes. If left untreated, the patient may ultimately have complete loss of vision. While there are effective therapies for diabetic retinopathy (DR), it is crucial to detect the condition at an early stage and continuously monitor patients with diabetes. Certain physical examinations, such as visual acuity tests, pupil dilation, and optical coherence tomography can also be employed for the diagnosis of diabetic retinopathy, but with a lengthier duration. The primary objective of this study is to employ the optimised back propagation-based deep residual learning network algorithm (Op-BPDRLN) using extracted characteristics from diverse retinal image outputs. The aim is to enhance the accuracy of diagnosing diabetes retinopathy disease while reducing the detection time. The evaluation of the output of the suggested classification model is conducted based on metrics such as accuracy, error, precision, and recall, with the aim of determining its superiority. The superiority of the proposed method over the existing approach has been substantiated through the analysis of simulation data.