Published Paper Details:

Diabetic retinopathy (DR) is a significant consequence of diabetic condition (Hyperglycemia) and a primary factor in visual deficiency globally. Prompt identification and timely intervention play a pivotal role in averting vision impairment in individuals with diabetes. Lately, progress in deep learning techniques, notably Convolutional Neural Networks (CNNs), has showcased significant successes in various medical image analysis tasks, including the detection of diabetic retinopathy. This paper introduces an innovative method for automatically detecting diabetic retinopathy utilizing CNNs. Our proposed method involves preprocessing retinal fundus images to enhancing image contrast and reducing noise, followed by feature selection using a pretrained CNN architecture. The extracted features are then fed into a classification model for the recognition of Retinal disease in diabetes. We utilize a voluminous dataset of annotated retinal visuals to train & validate our CNN-based detection system, ensuring robust performance across diverse clinical scenarios. Experimenting the outcomes that showcase the efficiency of our approach in precisely achieving the desired results in detecting diabetic retinopathy, achieving cutting edge performance in sensitivity, specificity, and overall accuracy. Moreover, the proposed method exhibits robustness to variations in image quality and pathological characteristics, making it suitable for real-world clinical applications. In summary, our research underscores the promise of leveraging deep learning, particularly CNNs, as an invaluable resource in the prompt identification and treatment of diabetic retinopathy. Our proposed model offers potential for seamless integration into current healthcare infrastructure, enabling timely detection and intervention strategies to mitigate vision loss in individuals with diabetes.