Table 4 Comparison results before and after data augmentation of five CNNs
From: Automatic freezing-tolerant rapeseed material recognition using UAV images and deep learning
Dataset | Models | Accurac y (%) | Recall (%) | Precision (%) | F-score (%) |
|---|---|---|---|---|---|
Original dataset | AlexNet | 92.63 | 73.40 | 80.23 | 76.67 |
VGGNet16 | 93.16 | 73.40 | 83.13 | 77.97 | |
GoogLeNet | 92.45 | 67.02 | 84.00 | 74.56 | |
ResNet18 | 92.98 | 69.15 | 85.53 | 76.47 | |
ResNet50 | 93.33 | 79.79 | 79.79 | 79.79 | |
Augmented dataset | AlexNet | 92.98 | 69.15 | 85.53 | 76.47 |
VGGNet16 | 93.68 | 76.60 | 83.72 | 80.00 | |
GoogLeNet | 92.80 | 71.28 | 82.72 | 76.57 | |
ResNet18 | 93.33 | 70.21 | 86.84 | 77.65 | |
ResNet50 | 93.68 | 79.79 | 81.52 | 80.65 |