Table 2.

CenterNet, YOLOv5, YOLOv7, and faster RCNN training model evaluation for the detection of defective seedlings.

Model	Precision	Recall	mAP @0.5
CenterNet	-	0.850	0.812
YOLOv5	0.936	0.990	0.973
YOLOv7	0.937	0.989	0.966
Faster R-CNN	0.853	1.000	0.982