Hybrid CNN+Transformer for Diabetic Retinopathy Recognition and Grading

Keyphrases

• CNN-Transformer 100%
• Diabetic Retinopathy Grading 100%
• Diabetic Retinopathy Screening 100%
• Grading System 66%
• Messidor 66%
• Performance Improvement 33%
• Diabetic Retinopathy 33%
• Deep Convolutional Neural Network (deep CNN) 33%
• Early Treatment 33%
• Training Data 33%
• Ablation Study 33%
• Art Performance 33%
• Transformer Encoder 33%
• Clinical Application 33%
• Transformer 33%
• Detection System 33%
• CNN-based 33%
• Feature Encoding 33%
• Feedforward Network 33%
• Benchmark Dataset 33%
• Feature Extractor 33%
• Early Diagnosis 33%
• Small Dataset 33%
• Recognition System 33%
• EfficientNet 33%
• Multi-head Attention 33%
• Global Dependencies 33%
• Spatial Modules 33%
• APTOS 33%
• EyePACS 33%
• Computer-aided System 33%
• Large-scale Training 33%
• Causes of Blindness 33%

Computer Science

• Convolutional Neural Network 100%
• Art Performance 25%
• Feedforward Network 25%
• Good Performance 25%
• Training Dataset 25%
• Clinical Application 25%

Engineering

• Convolutional Neural Network 100%
• Clinical Application 25%
• Feedforward 25%
• Extractor 25%
• Good Performance 25%