Enhancing Diabetic Retinopathy Detection with an Innovative ACSE-CLF Algorithm: A Deep Learning Approach

Diabetic Retinopathy (DR) is an increasingly prevalent consequence of diabetes that can cause vision impairment if left untreated. Early identification is the key to effective DR therapy, hence it is crucial to identify DR at an early stage in order to achieve the best results. Conventional screening techniques, however, demand specialised equipment, are imprecise, time-consuming, and inaccurate. An novel Attention Convolutional Squeeze Excitation-based Chaotic Levy Flight (ACSE-CLF) technique is thus proposed for the precise identification of DR. This study emphasises possible impacts associated with using the proposed ACSE-CLF algorithm with data pre-processing, feature extraction, and classification for recognising DR. The proposed ACSE-CLF method differentiates between DR and healthy images, improving screening precision. The performance has been evaluated using Kaggle datasets, which also included patient retinal images. These retinal images served as the proposed model’s input parameters, and achieved an accuracy rate of 98.7% for DR recognition.