Table 2.

Comparison of precision indexes for each algorithm training.

Evaluation metrics	YOLOv4	YOLOv5	YOLOX	YOLOv7	EfficientDet	Faster R-CNN	CenterNet
Precision	93.39%	90.09%	89.34%	89.39%	88.84%	62.79%	92.91%
Recall	90.99%	89.70%	93.56%	93.99%	92.27%	97.21%	87.12%
F1 score	0.92	0.89	0.91	0.92	0.91	0.76	0.90
mAP	94.30%	94.21%	95.47%	94.59%	93.74%	93.45%	93.53%