Table 1.

Test performance on various models.

Model	Accuracy	Recall	Precision	AUC
Custom model (ResNet50 Base)	0.7866	0.6873	0.9615	0.9023
With contrast enhancement	0.8190	0.7423	0.9600	0.9020
Custom model (ResNet101 Base)	0.7845	0.7113	0.9283	0.8823
With contrast enhancement	0.7931	0.7354	0.9185	0.8834
Custom model (ResNet152 Base)	0.7672	0.6667	0.9463	0.8664
With contrast enhancement	0.7672	0.6667	0.9463	0.8724
ResNet50 Base (no additional layers or training)	0.3685	0.0103	0.3750	0.2815
With contrast enhancement	0.3685	0.0103	0.3750	0.2812