Table 9 Comparison of performance evaluation indicators of different object detection models on the test set of this study.
From: Object detection model design for tiny road surface damage
Models | P(%) | R(%) | mAP50(%) | mAP50:95(%) | F1 | Params(M) | GFLOPs | Latency(ms) |
|---|---|---|---|---|---|---|---|---|
YOLOv8n | 63.1 | 53.9 | 57.5 | 29.4 | 58.1 | 5.97 | 8.1 | 4.4 |
YOLOv8s | 69.7 | 61.3 | 65.3 | 34.9 | 65.2 | 21.4 | 28.4 | 5.1 |
YOLOv8m | 72.4 | 63.7 | 68.2 | 37.1 | 67.8 | 49.6 | 78.7 | 7.0 |
YOLOv8l | 72.1 | 63.9 | 67.9 | 36.9 | 67.8 | 83.6 | 164.8 | 10.4 |
YOLOv8x | 73.3 | 65.3 | 69.0 | 37.4 | 69.1 | 130 | 257.4 | 14.7 |
YOLOv6n | 54.8 | 45.8 | 47.3 | 24.2 | 49.9 | 8.3 | 11.8 | 3.8 |
YOLOv6s | 58.6 | 52.6 | 54.6 | 28.3 | 55.4 | 31.3 | 44.0 | 4.6 |
YOLOv6m | 63.9 | 52.0 | 56.0 | 30.0 | 57.3 | 99.5 | 161.1 | 8.9 |
YOLOv6l | 57.7 | 44.0 | 48.0 | 24.8 | 49.9 | 211 | 391.2 | 16.1 |
YOLOv5n | 60.6 | 52.9 | 56.0 | 28.3 | 56.5 | 5.04 | 7.1 | 3.8 |
YOLOv5s | 70.1 | 59.9 | 64.7 | 34.2 | 64.6 | 17.6 | 23.8 | 4.3 |
YOLOv5m | 71.2 | 63.6 | 68.4 | 36.4 | 67.2 | 48.1 | 64 | 6.7 |
YOLOv5l | 72.3 | 64.5 | 68.3 | 37.4 | 68.2 | 101 | 134.7 | 9.3 |
YOLOv5x | 73.8 | 64.7 | 69.5 | 38.0 | 69.0 | 185 | 246 | 15.8 |
YOLOv9 | 71.2 | 62.3 | 67.7 | 37.6 | 66.5 | 240 | 116.8 | 13.3 |
Ours | 72.4 | 68.2 | 70.8 | 38.0 | 70.2 | 16.5 | 30.9 | 4.5 |
