Efficient Detection of Diabetic Retinopathy through Deep Learning
Efficient Detection of Diabetic Retinopathy through Deep Learning
IEEE BASE PAPER TITLE:
A Lesion-Based Diabetic Retinopathy Detection Through Hybrid Deep Learning Model
IEEE BASE PAPER ABSTRACT:
Diabetic retinopathy (DR) can be defined as visual impairment caused by prolonged diabetes affecting the blood vessels in the retina. Globally, it stands as the primary contributor to blindness, impacting approximately 191 million individuals. While prior research has addressed DR classification using retinal fundus images, existing methods often focus on isolated lesion detection, lacking a comprehensive framework for the simultaneous identification of all lesions. Previous studies concentrated on early-stage features like exudates, aneurysms, hemorrhages, and blood vessels, sidelining severe-stage lesions such as cotton wool spots, venous beading, very severe intraretinal microvascular abnormalities (IRMA), diffuse intraretinal hemorrhages, capillary degeneration, highly activated microglia, and retinal pigment epithelium (RPE) damage. In this study, a deep learning approach is proposed to classify DR fundus images by severity levels, utilizing GoogleNet and ResNet models based on adaptive particle swarm optimizer (APSO), for enhanced feature extraction. The extracted features from the hybrid model are further used with different machine learning models like random forest, support vector machine, decision tree, and linear regression models. Experimental results showcased the proposed hybrid framework outperforming advanced approaches with a remarkable 94% accuracy on the benchmark dataset. This method demonstrates potential enhancements in precision, recall, accuracy, and F1 score for different DR severity levels.
PROJECT OUTPUT VIDEO:
ALGORITHM / MODEL USED:
MobileNetV2 Architecture and DenseNet201 Architecture.
OUR PROPOSED PROJECT ABSTRACT:
Diabetic retinopathy (DR) is a leading cause of blindness globally, necessitating timely and accurate diagnosis to prevent vision loss. This project presents an efficient detection system for diabetic retinopathy using deep learning techniques, integrating modern web technologies to provide a user-friendly interface for medical professionals.
The detection system is developed using Python for backend processing, with Flask serving as the web framework. The frontend interface is built with HTML, CSS, and JavaScript, ensuring an intuitive and responsive user experience.
The dataset utilized comprises 1,812 retinal images categorized into five distinct groups: ‘normal’, ‘mild’, ‘moderate’, ‘severe’, and ‘proliferative’ diabetic retinopathy. This diverse dataset ensures comprehensive training and validation of the models, covering the full spectrum of DR severity.
Two state-of-the-art convolutional neural network architectures, MobileNetV2 and DenseNet201, were employed to classify retinal images into five categories: ‘normal’, ‘mild’, ‘moderate’, ‘severe’, and ‘proliferative’ diabetic retinopathy. The dataset comprises 1,812 retinal images, meticulously labeled to facilitate accurate model training and evaluation. MobileNetV2 achieved a remarkable training accuracy of 97.00% and a validation accuracy of 99.00%, demonstrating its superior performance and reliability in detecting diabetic retinopathy. In contrast, DenseNet201 achieved a training accuracy of 90.00% and a validation accuracy of 88.00%, providing a comparative benchmark for model performance.
Our system’s high accuracy rates, particularly with the MobileNetV2 architecture, highlight its potential for real-world application in clinical settings. The integration with a web-based interface facilitates easy deployment and accessibility, promoting widespread adoption in the medical community. This project signifies a significant step forward in leveraging deep learning for the early and efficient detection of diabetic retinopathy, ultimately contributing to improved patient outcomes.
SYSTEM REQUIREMENTS:
HARDWARE REQUIREMENTS:
- System : Pentium i3 Processor.
- Hard Disk : 500 GB.
- Monitor : 15’’ LED.
- Input Devices : Keyboard, Mouse.
- Ram : 8 GB.
SOFTWARE REQUIREMENTS:
- Operating System : Windows 10 / 11.
- Coding Language : Python 3.10.9.
- Web Framework : Flask.
- Frontend : HTML, CSS, JavaScript.
REFERENCE:
AYESHA JABBAR, HANNAN BIN LIAQAT, AFTAB AKRAM, MUHAMMAD USMAN SANA, IRMA DOMÍNGUEZ AZPÍROZ, ISABEL DE LA TORRE DIEZ, AND IMRAN ASHRAF, “A Lesion-Based Diabetic Retinopathy Detection Through Hybrid Deep Learning Model”, IEEE Access, 2024.