Table 3 Performance comparison of different methods on the DVTOD dataset.
From: Cross-modal edge-enhanced detector for UAV-based multispectral object detection
Methods | Person | Car | Bicycle | mAP@50 | Modality |
|---|---|---|---|---|---|
YOLOv348 | 25.1 | 53.8 | 24.6 | 34.5 | Â |
YOLOv549 | 23.3 | 50.2 | 26.7 | 33.4 | Â |
YOLOv650 | 31.0 | 52.7 | 30.5 | 38.1 | RGB |
YOLOv751 | 26.5 | 45.0 | 20.4 | 30.6 | Â |
YOLOv8 | 28.4 | 52.7 | 24.7 | 35.3 | Â |
YOLOv348 | 87.1 | 76.1 | 84.8 | 82.7 | Â |
YOLOv549 | 86.2 | 75.9 | 86.2 | 82.8 | Â |
YOLOv650 | 86.2 | 78.3 | 88.8 | 84.4 | IR |
YOLOv751 | 80.2 | 68.7 | 84.5 | 77.8 | Â |
YOLOv8 | 86.2 | 75.6 | 88.1 | 83.3 | Â |
YOLOv5 + Add | 88.8 | 74.3 | 74.6 | 79.2 |  |
YOLOv5 + CMX52 | 88.9 | 75.9 | 79.6 | 81.6 | RGB + IR |
CFT53 | 88.9 | 78.0 | 81.3 | 82.7 | |
CMEE-Det | 90.8 | 80.2 | 81.4 | 84.4 | Â |
