Table 7 Performances of the proposed FRYOLO model in detecting different fresh and defective fruits.
From: An optimized lightweight real-time detection network model for IoT embedded devices
Type | Precision (%) | Recall (%) | mAP (%) |
|---|---|---|---|
Fresh apple | 94.9 | 92.7 | 97.2 |
Fresh mango | 80.6 | 87.1 | 89.8 |
Fresh orange | 95.5 | 72.8 | 94.7 |
Fresh pear | 91.2 | 94.3 | 96.5 |
Fresh banana | 78.1 | 100 | 97.9 |
Defective apple | 85.2 | 81.6 | 88.0 |
Defective mango | 76.9 | 82.4 | 82.9 |
Defective orange | 82.6 | 89.0 | 93.9 |
Defective pear | 81.6 | 95.9 | 99.5 |
