Research on grading detection methods for diabetic retinopathy based on deep learning.

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Diabetic retinopathy is a leading cause of vision loss.
• Early detection and intervention are crucial for preventing severe vision impairment.
• Current screening methods can be time-consuming and require specialized expertise.

Purpose of the Study:

• To develop a deep learning model for the early screening of diabetic retinopathy.
• To predict the presence of diabetic retinopathy from fundus images.
• To provide interpretable justifications for the model's predictions.

Main Methods:

• A Vision Transformer-based deep learning architecture was employed.
• The model was trained using the publicly available EyePACS dataset.
• The model predicts diabetic retinopathy and identifies affected regions in fundus images.

Main Results:

• The model achieved a detection accuracy of approximately 0.88.
• The model demonstrated comparable performance to existing methods in identifying affected regions.
• Interpretability was provided by highlighting areas of concern in the images.

Conclusions:

• A deep learning-based method for diabetic retinopathy detection was successfully implemented.
• The model provides interpretable justifications, aiding clinicians in diagnosis.
• This approach facilitates early screening and management of diabetic retinopathy.