Table 3.

Comparison of the confusion matrix results and its parameters for the four models in the test set.

Models	Accuracy	Sensitivity	Specificity	AUC	Mean IoU	F1 score	p-value
Faster RCNN (ResNet 50)	0.88	0.94	0.71	0.91	0.72	0.92	0.001
Faster RCNN (Inception V3)	0.83	0.91	0.59	0.85	0.74	0.89	0.001
YOLO V7 (ResNet 50)	0.86	0.94	0.61	0.86	0.68	0.91	0.001
YOLO V7 (Inception V3)	0.83	0.91	0.60	0.83	0.69	0.89	0.001

Mean IoU, mean intersection over union; AUC, area under curve; YOLO V7, You Only Look Once Version 7; RCNN, Region-Based Convolutional Neural Network.