Table 2.

Performance of ResNet-50 classification on training, validation, and test sets. Accuracy, precision, and recall are used for evolution.

	Accuracy	Precision	Recall
Training	0.9632	0.9644	0.9632
Validation	0.7879	0.7890	0.7879
Test	0.7576	0.7540	0.7576
External test set	0.6779	0.6802	0.6639