Table 2.

Overview on all metrics of the three models YOLOv5n, YOLOv5s, and YOLOv7t.

	Number parameters (m)	CPU inference [sec/img]	GPU inference [sec/img]	Conf. & IoU Threshold	Overall accuracy	Precision	Recall	False positive rate	IoU
YOLOv5n	1.9	0.1893	0.0552	0.2 & 0.5	0.9450	0.8427	0.8305	0.0309	0.7831
YOLOv5s	7.2	0.4833	0.0554	0.3 & 0.6	0.9624	0.9088	0.8692	0.0209	0.8052
YOLOv7t	6.2	0.4047	0.0734	0.1 & 0.3	0.9508	0.8449	0.8907	0.0283	0.8026