Table 4 Comparative model results.

From: A deep learning-based approach for axle counter in free-flow tolling systems

Model

Precision

Recall

mAP @0.50

mAP @0.50:0.95

FLOPs

YOLOv5m

0.994

0.982

0.994

0.800

49.0

YOLOv6m

0.990

0.832

0.989

0.792

82.2

YOLOv7x

0.979

0.923

0.975

0.665

189.9

YOLOv8m

0.992

0.980

0.994

0.820

78.9

  1. Best performance values are in bold.