Table 7 Results of comparison with other algorithms on the TT100K dataset.
From: YOLO-BS: a traffic sign detection algorithm based on YOLOv8
Model | GFLOPs | P/% | R/% | mAP50/% | mAP50–95/% | FPS |
|---|---|---|---|---|---|---|
Faster R-CNN | 134 | 58.8 | 52.3 | 55.7 | 42.4 | 45 |
Mask R-CNN | 187 | 55.0 | 52.8 | 53.9 | 40.5 | 37 |
TOOD | 125 | 61.3 | 53.7 | 71.4 | 51.5 | 47 |
YOLOv5 | 16.1 | 82.5 | 77.1 | 83.1 | 63.4 | 39 |
YOLOv7 | 103.9 | 52.1 | 55.4 | 52.9 | 39.4 | 65 |
YOLOv9 | 314.0 | 87.3 | 80.2 | 87.9 | 68.8 | 31 |
YOLOv10 | 24.6 | 84.9 | 72.3 | 82.4 | 64.6 | 83 |
Wang et al. 18 | 17.9 | – | – | 65.1 | – | 95 |
Wang et al. 13 | – | – | 64.5 | 52 | – | 87 |
Li et al. 19 | – | – | – | 89.7 | – | 65 |
Gong et al.20 | – | – | – | 77.3 | – | 33 |
Ours(YOLO-BS) | 55.2 | 87.9 | 80.5 | 90.1 | 70.3 | 78 |
