Table 4 Comparison of detection results after replacing the loss function with YOLOX.
From: Tomato ripeness and stem recognition based on improved YOLOX
Model | Precision/% | Recall/% | F1/% | mAP/% | AP/% | |||
|---|---|---|---|---|---|---|---|---|
Red | Half | Green | Stem | |||||
YOLOX | 83.39 | 87.51 | 85.25 | 91.00 | 97.50 | 85.64 | 95.81 | 85.07 |
YOLOX-DIoU | 81.53 | 85.16 | 83.25 | 91.07 | 97.36 | 85.37 | 95.85 | 85.67 |
YOLOX-CIoU | 84.51 | 86.06 | 86.00 | 91.35 | 97.59 | 86.22 | 96.26 | 85.33 |
YOLOX-GIoU | 84.72 | 88.11 | 86.25 | 91.44 | 97.33 | 86.65 | 95.61 | 86.18 |
