Table 3 Comparative experimental results of YOLOv8-RTDAV on the KITTI dataset.
From: An algorithm for road target detection of autonomous vehicles based on improved YOLOv8
Model | Precision/% | Recall/% | mAP@50/% | Parameters | GFLOPs | Model size/MB | FPS |
|---|---|---|---|---|---|---|---|
YOLOv3-tiny | 0.829 | 0.648 | 0.725 | 12,129,206 | 18.9 | 23.2 | 276 |
YOLOv5n | 0.849 | 0.7 | 0.786 | 2,507,401 | 7.1 | 5.04 | 127 |
YOLOv6n | 0.853 | 0.671 | 0.76 | 4,234,425 | 11.8 | 8.3 | 189 |
YOLOv8n | 0.88 | 0.763 | 0.852 | 3,009,977 | 8.2 | 5.97 | 148 |
YOLOv9t | 0.821 | 0.693 | 0.775 | 2,005,177 | 7.8 | 4.43 | 48 |
YOLOv10n | 0.872 | 0.721 | 0.802 | 2,702,146 | 8.3 | 5.5 | 103 |
YOLOv8-RTDAV | 0.869 | 0.811 | 0.88 | 2,811,620 | 11.8 | 5.78 | 102 |
