Table 5 The detection results of the selected detectors on the proposed dataset.
Model | mAP (%) | Precision (%) | Recall (%) | F1-score (%) | FLOPS | Parameters |
|---|---|---|---|---|---|---|
CenterNet | 82.67 | 93.59 | 58.33 | 71.86 | 70.22G | 32.67 MB |
SSD | 91.48 | 86.66 | 86.87 | 86.77 | 72.11G | 38.85 MB |
EfficientDet-D0 | 87.87 | 88.63 | 81.78 | 85.07 | 5.45G | 3.90 MB |
Faster R-CNN | 89.82 | 67.65 | 93.29 | 78.43 | 372.52G | 139.02 MB |
RetinaNet | 87.10 | 86.96 | 80.16 | 83.43 | 236.06G | 57.69 MB |
YOLOv3 | 87.77 | 88.91 | 82.04 | 85.34 | 66.43G | 62.14 MB |
YOLOv4 | 82.18 | 83.95 | 76.27 | 79.92 | 60.78G | 64.55 MB |
YOLOv5-X | 88.17 | 88.10 | 84.92 | 86.48 | 220.25G | 88.01 MB |
YOLOv7 | 90.97 | 89.47 | 88.01 | 88.73 | 107.07G | 37.81 MB |
YOLOv7-tiny | 81.82 | 87.16 | 72.31 | 79.05 | 14.16G | 6.32 MB |
YOLOv8-N | 88.96 | 89.95 | 85.12 | 87.47 | 10.46G | 3.50 MB |
YOLOX-N | 82.53 | 85.73 | 69.71 | 76.90 | 2.67G | 0.92 MB |
