Early Assessment of Diabetic Retinopathy by Detecting Microaneurysm with Deep Neural Network

Abstract

Objectives: This study aims to detect early signs of diabetic retinopathy, also known as microaneurysm (MA). Early detection of MA can help in diagnosing diabetic retinopathy effectively. Methods: To achieve this objective a method is proposed which is based on a deep learning model that incorporates transfer learning. The dataset used in this study is E-Ophtha which consists of 381 high-quality images. The proposed model consists of three steps which are preprocessing, feature extraction and classification. The method uses CLAHE to enhance the details of the fundus image. The feature extraction step and the classification are done using a deep learning algorithm. The proposed method is compared by using parameters such as Recall, Precision, and accuracy. Findings: The accuracy achieved by the model is about 98.82%-100%, a recall/sensitivity of 100%, and precision is also 100% which is close to the state of the art. This implies that deep learning methods are a good fit for identifying microaneurysms and ultimately detecting diabetic retinopathy. Novelty: Early detection and diagnosis of diabetic retinopathy can be achieved with this approach, and appropriate medication can impede the disease's development. Keywords: Diabetic Retinopathy, PDR, NPDR, Image Processing, Deep Learning