Table 4 Performance comparison of baseline models and their enhanced versions using our proposed refinement framework on the EyePACS test dataset.

From: Diabetic retinopathy classification using a multi-attention residual refinement architecture

Methods

Baseline models

Enhanced models (+ours)

Sens.

Spec.

Kappa

AUC

Sens.

Spec.

Kappa

AUC

MobileNet V327

0.76

0.74

0.76

0.81

0.78

0.77

0.79

0.84

ResNet5029

0.78

0.77

0.78

0.84

0.80

0.81

0.82

0.89

ResNet10129

0.79

0.78

0.79

0.85

0.81

0.82

0.83

0.90

DenseNet12130

0.80

0.78

0.78

0.86

0.82

0.81

0.82

0.91

EfficientNet-B0

0.77

0.76

0.77

0.83

0.79

0.79

0.80

0.86

  1. The enhancement includes space-to-depth preprocessing, expansion layers, SE blocks, and multi-attention mechanism (\(H=4\)).
Back to article page