Enhanced Deep Learning Model for Classification of Retinal Optical Coherence Tomography Images

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Retinal optical coherence tomography (OCT) is crucial for diagnosing eye conditions and monitoring systemic diseases like diabetes.
• Accurate diagnosis and automated analysis of OCT images are essential for effective clinical practice.
• Existing models require enhancement for improved specificity and performance in retinal image classification.

Purpose of the Study:

• To propose an enhanced optical coherence tomography (EOCT) model for classifying retinal OCT images.
• To improve the performance of automated retinal image analysis using deep learning.
• To achieve higher accuracy in diagnosing various retinal conditions.

Main Methods:

• Development of an enhanced optical coherence tomography (EOCT) model.
• Utilizing a modified ResNet (50) architecture combined with random forest algorithms for classification.
• Employing the Adam optimizer to enhance the efficiency of the ResNet (50) model during training.

Main Results:

• The EOCT model achieved high performance metrics, including sensitivity (0.9836), specificity (0.9615), and accuracy (0.9747).
• The model demonstrated superior efficiency compared to pre-trained models like VGG (16).
• Key performance indicators such as precision and Matthew's correlation coefficient were also notably high, indicating robust classification capabilities.

Conclusions:

• The proposed EOCT model offers a significant advancement in automated retinal OCT image classification.
• This enhanced model holds promise for improving diagnostic accuracy and patient outcomes in ophthalmology.
• Further research can explore the integration of EOCT in clinical workflows for real-time disease detection.