Table 7 Comparison of detection results using different network models
From: Early recognition of tomato gray leaf spot disease based on MobileNetv2-YOLOv3 model
Test set | Network models | F1 score/Â % | Average precision/Â % |
|---|---|---|---|
Sufficient light without leaf shelter | GIoU + MobileNetv2-YOLOv3 | 94.13 | 92.53 |
SSD | 92.01 | 89.18 | |
Faster-RCNN | 92.45 | 89.42 | |
Sufficient light with leaf shelter | GIoU + MobileNetv2-YOLOv3 | 93.22 | 91.01 |
SSD | 91.44 | 88.52 | |
Faster-RCNN | 92.12 | 92.13 | |
Insufficient light without leaf shelter | GIoU + MobileNetv2-YOLOv3 | 91.32 | 90.07 |
SSD | 89.67 | 87.96 | |
Faster-RCNN | 90.01 | 88.33 | |
Insufficient light with leaf shelter | GIoU + MobileNetv2-YOLOv3 | 90.61 | 90.02 |
SSD | 88.55 | 86.52 | |
Faster-RCNN | 89.77 | 87.61 |