A Novel Approach to Identify Diabetic Macular Edema Using a Minimal CNN Model

DME ranks among the primary factors behind the decline in eyesight among individuals with diabetes, and Prompt identification can aid in avoiding irreversible damage. However, current DME detection methods rely on complex and computationally intensive models, making them unfit for widespread use. This study aims to develop a novel Minimal Convolutional Neural Network (MCNN) framework for the early identification of Diabetic Macular Edema (DME) in medical Photographs. The proposed minimal CNN model uses a smaller number of convolutional layers and filters than conventional models while achieving comparable or better Accomplishment. The training of the model involves a substantial collection of retinal images from both patients with and without DME. The outcomes demonstrate that the proposed work can accurately detect DME with high sensitivity and specificity, providing a promising tool for early diagnosis and effective management of the condition. The novel approach presented in this study achieved an accuracy of 91.14%.