Table 5 Comparison experiments of different models on the VisDrone dataset.(CF-YOLO is an enhanced version based on YOLOv11n).
From: CF-YOLO for small target detection in drone imagery based on YOLOv11 algorithm
Model | P | R | F1 | mAP50 | mAP50:95 | Para(M) | GFLOPs |
|---|---|---|---|---|---|---|---|
YOLOv3-tiny | 39.1 | 24.3 | 22.5 | 23.6 | 13.2 | 14.3 | 9.52 |
YOLOv5n | 44.5 | 33.2 | 38.0 | 32.9 | 19.1 | 5.8 | 2.18 |
YOLOv5s | 51.1 | 38.1 | 43.7 | 39.3 | 23.4 | 18.8 | 7.81 |
YOLOv5l | 50.7 | 38.6 | 43.9 | 41.4 | 24.6 | 107.8 | 46.15 |
YOLOv6s | 40.3 | 30.5 | 30.2 | 17.7 | 11.5 | 4.15 | – |
YOLOv7-tiny | 47.6 | 37.3 | 41.8 | 35.8 | 18.8 | 13.3 | 6.04 |
YOLOv8n | 45 | 33 | 38.1 | 33.1 | 19.2 | 6.8 | 2.68 |
YOLOv8s | 50.7 | 37.9 | 43.3 | 39.1 | 23.4 | 23. | 49.83 |
YOLOv8m | 53.3 | 41.1 | 46.4 | 42.5 | 26 | 67.5 | 23.2 |
YOLOv9s | 52 | 38 | 43.9 | 39.4 | 23.8 | 22.1 | 61.9 |
YOLOv10n | 45.0 | 34.5 | 39.1 | 34.5 | 19.9 | 6.5 | 2.26 |
YOLOv10s | 52.7 | 38 | 44.0 | 39.8 | 23.8 | 21.4 | 7.22 |
YOLOv10m | 55.1 | 42.1 | 47.7 | 44.2 | 26.9 | 58.9 | 15.31 |
YOLOv11n | 42.8 | 33.1 | 37.3 | 32.2 | 18.6 | 6.3 | 2.58 |
YOLOv11s | 49.9 | 38.7 | 43.5 | 39.4 | 23.6 | 21.3 | 9.41 |
YOLOv11m | 55.7 | 42.5 | 48.2 | 44.1 | 27.2 | 67.7 | 20.03 |
RT-DETR(r18) | 57.2 | 40 | 47.1 | 41.4 | 25.1 | 57 | 20 |
EL-YOLO39 | 48.8 | 40.3 | 43 | 42.9 | 24.8 | 6.7 | 1.08 |
EBC-YOLO56 | 55.3 | 42.0 | 47.7 | 44.3 | 26.7 | 35.5 | 10.2 |
YOLOv8-QSD15 | 44.2 | 34.2 | 38.6 | 34.6 | 16.8 | – | – |
EdgeYOLO57 | – | – | – | 44.8 | 26.4 | – | 40.5 |
Drone-YOLO58 | – | – | – | 42.8 | 25.6 | – | 5.35 |
CF-YOLO | 52.8 | 43.4 | 47.7 | 44.9 | 27.5 | 23.9 | 3.77 |
