Table 3 Accuracy performance analysis of methodologies
From: A hybrid CNN–SVM classifier for weed recognition in winter rape field
Method | VGG-SVM | VGG [35] | ANN [16] | SVM [10] | ResNet [36] | Trong et al. [37] |
|---|---|---|---|---|---|---|
Test times | Accuracy (%) | Accuracy (%) | Accuracy (%) | Accuracy (%) | Accuracy (%) | Accuracy (%) |
1 | 92.3 | 90.3 | 89 | 87.7 | 89.7 | 91.3 |
2 | 92.3 | 89.7 | 88 | 85.7 | 90.3 | 91 |
3 | 91.3 | 91.3 | 88.3 | 88 | 90 | 90.3 |
4 | 92 | 90.7 | 88.3 | 86.7 | 88.7 | 90 |
5 | 92.7 | 90.7 | 86 | 87.7 | 89.3 | 91.3 |
Mean | 92.1 | 90.5 | 87.9 | 87.2 | 89.6 | 90.8 |
Std | 0.52 | 0.59 | 1.13 | 0.87 | 0.62 | 0.60 |