Diagnosis of Retinal Diseases Based on Bayesian Optimization Deep Learning Network Using Optical Coherence Tomography...

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Retinal abnormalities pose a significant public health risk, potentially leading to vision impairment and blindness.
• Early and accurate detection of retinal diseases is crucial for effective treatment and prevention of vision loss.
• Current diagnostic methods may not always provide the necessary precision or early detection.

Purpose of the Study:

• To investigate the efficacy of transfer learning using pretrained Convolutional Neural Network (CNN) models for detecting retinal diseases.
• To classify seven different retinal diseases using Optical Coherence Tomography (OCT) images.
• To compare the performance of various CNN models and identify the most accurate approach.

Main Methods:

• Utilized pretrained CNN models: VGG16, DenseNet201, InceptionV3, and Xception.
• Applied transfer learning to classify retinal diseases from a dataset of OCT images.
• Employed Bayesian optimization for hyperparameter tuning and image augmentation to enhance model generalization.

Main Results:

• DenseNet201 achieved an accuracy exceeding 99% in classifying retinal diseases from OCT images.
• The developed models demonstrated high accuracy in differentiating between healthy and diseased retinal images.
• A comparative analysis showed DenseNet201 outperformed other evaluated CNN models.

Conclusions:

• Transfer learning with CNNs, particularly DenseNet201, offers a highly accurate automated approach for retinal disease detection.
• This method provides a significant advancement over existing techniques, enabling classification of multiple retinal diseases.
• The findings support the potential of AI-driven tools for early and precise diagnosis of vision-threatening retinal conditions.