Research Article
[Retracted] Cataract Disease Detection by Using Transfer Learning-Based Intelligent Methods
Table 3
Performance score of DensNet121, Xception, InceptionV3, and InceptionResNetV2.
| Model | Dataset | Accuracy | Loss | Sensitivity | Specificity | Images | Precision | Recall | F1 score |
| DenseNet121 | Test | 0.9541 | 0.2313 | 0.9230 | 0.9842 | Normal | 0.98 | 0.92 | 0.95 | Cataract | 0.93 | 0.98 | 0.96 | Xception | Test | 0.9771 | 0.0719 | 0.9792 | 0.9754 | Normal | 0.97 | 0.98 | 0.97 | Cataract | 0.98 | 0.98 | 0.98 | InceptionV3 | Test | 0.9771 | 0.1223 | 0.9504 | 1.00 | Normal | 1.00 | 0.95 | 0.97 | Cataract | 0.96 | 1.00 | 0.98 | InceptionResNetV2 | Test | 0.9817 | 0.0622 | 0.9655 | 1.00 | Normal | 1.00 | 0.97 | 0.98 | Cataract | 0.96 | 1.00 | 0.98 |
|
|