Research Article
Diagnosing Diabetic Retinopathy in OCTA Images Based on Multilevel Information Fusion Using a Deep Learning Framework
Table 5
The classification results on another OCTA dataset.
| Model | Accuracy | Sensitivity | Specificity |
| EfficientNet | 68.2% () | 56.7% () | 64.8% () | MobileNet v2 | 66.2% () | 68.9% () | 53.7% () | ResNet50 | 75.8% () | 62.2% () | 69.4% () | ShuffleNet v2 | 73.8% () | 62.2% () | 71.3% () | SqueezeNet | 57.0% () | 0.9% () | 98.2% () | Ours | 76.0% () | 76.2% () | 75.9% () |
|
|