Table 7 BioS-Wheat + GWHD dataset–Results.
From: Transfer learning models for wheat ear detection on multi-source dataset
Model | mAP@50 | Training duration | Inference time |
|---|---|---|---|
F-RCNN R 50 | 81.48 ± 2.31 | 1h 49min 34s | 65 ms |
F-RCNN R 101 | 72.36 ± 0.69 | 2h 22min 12s | 27 ms |
F-RCNN X 101 32x8d | 84.33 ± 0.37 | 4h 36min 21s | 116 ms |
RetinaNet R 101 | 87.79 ± 0.76 | 2h 17min 41s | 85 ms |
YOLOv8 | 88.00 ± 0.56 | 5h 34min 11s | 18 ms |
RT-DETR | 91.00 ± 0.41 | 9h 14min 42s | 137 ms |
