Table 1 Comparison of license plate detection algorithms on CCPD dataset.
From: License plate recognition system for complex scenarios based on improved YOLOv5s and LPRNet
Network | P | mAP50 | FPS | Params (M) | Model size (MB) | GFLOPs |
|---|---|---|---|---|---|---|
SSD | 0.883 | 0.860 | 47 | 23.90 | 45.80 | 24.4 |
Faster RCNN | 0.899 | 0.875 | 32 | 41.10 | 80.39 | 43.96 |
YOLOv7x | 0.980 | 0.993 | 86 | 70.78 | 134.99 | 188.9 |
KDNet | 0.988 | 0.991 | 125 | 39.59 | 75.5 | 107.8 |
YOLOv5s | 0.952 | 0.991 | 165 | 7.26 | 13.84 | 17.6 |
Ours | 0.989 | 0.994 | 147 | 7.49 | 14.28 | 18.1 |
