A Novel CNN architecture for detecting Diabetic Retinopathy

Abstract

Objective: Diabetic Retinopathy (DR) is the leading cause of preventable blindness in adults. It is the leading cause of vision loss, in people aged 50 and above. Global prevalence is projected to rise to 129.84 million by 2030 and 160.5 million by 2045. Effective management relies on regular screening and prompt intervention. With increasing demands, automated screening methods using deep learning techniques offer enhanced diagnostic accuracy and timely intervention. This study aimed to address the increasing prevalence of DR and the demand for automated screening methods, leveraging deep learning techniques to enhance diagnostic accuracy and facilitate timely intervention.

Methods: In this research, a new Convolutional Neural Network (CNN) architecture is tested with the goal of accurately detecting different stages of DR. A thorough assessment of innovative CNN design, is done using various metrics such as Accuracy, Precision, Recall, and F1 score.

Results: The Experimental findings indicate that the proposed CNN model surpasses existing research efforts, demonstrating its superiority in accurately predicting stages of DR. With impressive metrics such as 95% accuracy, 93% precision, 94% recall, and a 93% F1 score, this novel CNN model showcases its efficacy in precisely assessing the severity of DR and the integration of batch normalization and dropout layers into the architecture aids in mitigating overfitting, further enhancing the model's performance and generalization capabilities.

Conclusion: In conclusion, the comprehensive experimentation and evaluation have established the robustness and scalability of the model, offering promising prospects for improved DR screening in clinical settings. The results emphasize the transformative potential of innovative deep learning architectures in revolutionizing DR diagnosis and management. With these advancements, the future of DR care can be predicted to improve patient outcomes and preserve vision.