Table 3 Performance pairs of training results of images before and after enhancement on target detection model. Red indicates better results.
From: A hybrid zero-reference and dehazing network for joint low-light underground image enhancement
Model | Precision↑ | Recall↑ | AP (Average Precision)↑ | F1↑ |
|---|---|---|---|---|
YOLOv5(original) | 93.27 | 81.98 | 94.28 | 0.88 |
YOLOv5(enhanced) | 94.79 | 82.91 | 94.88 | 0.88 |
YOLOv8(original) | 98.02 | 89.19 | 97.91 | 0.93 |
YOLOv8(enhanced) | 95.69 | 94.87 | 97.92 | 0.95 |
SSD(original) | 93.07 | 80.34 | 88.63 | 0.86 |
SSD(enhanced) | 93.88 | 82.88 | 91.66 | 0.88 |
