Research Article
Hybrid Model Structure for Diabetic Retinopathy Classification
Table 1
Classification results from the basic models and the hybrid model structures.
| ā | Loss | Accuracy | Sensitivity | Specificity | Precision | F1 score |
| Basic models | EfficientNetB4 | CE | 0.8158 | 0.9442 | 0.7182 | 0.9027 | 0.9230 | E-CE | 0.8544 | 0.9736 | 0.7061 | 0.9017 | 0.9362 | EfficientNetB5 | CE | 0.7932 | 0.9254 | 0.6782 | 0.8884 | 0.9065 | E-CE | 0.8488 | 0.9809 | 0.6549 | 0.8872 | 0.9317 | NASNetLarge | CE | 0.7828 | 0.9151 | 0.7031 | 0.8951 | 0.9050 | E-CE | 0.8470 | 0.9845 | 0.6353 | 0.8820 | 0.9304 | InceptionResNetV2 | CE | 0.7888 | 0.9657 | 0.5177 | 0.8471 | 0.9025 | E-CE | 0.8502 | 0.9739 | 0.6963 | 0.8987 | 0.9348 | Xception | CE | 0.8100 | 0.9706 | 0.5742 | 0.8632 | 0.9138 | E-CE | 0.8476 | 0.9848 | 0.6217 | 0.8781 | 0.9284 |
| Hybrid model | Hybrid-model-a | CE | 0.8584 | 0.9877 | 0.6481 | 0.8860 | 0.9341 | Hybrid-model-f | CE | 0.8626 | 0.9652 | 0.7476 | 0.9137 | 0.9387 | Hybrid-model-c | CE | 0.8634 | 0.9706 | 0.7325 | 0.9094 | 0.9390 |
|
|
CE indicates cross-entropy loss function; E-CE indicates enhance cross-entropy loss function; and the bold values indicate the best results.
|