Table 5.
The accuracy (%) of the CNN models with image size = 64. The Accuracy Test Set was used for accuracy test for automatic identification of the five Korean softwoods
Species | LeNet | LeNet2 | LeNet3 | MiniVGGNet | MiniVGGNet2 | MiniVGGNet3 |
Cedar | 99.1 | 99.6 | 100.0 | 100.0 | 100.0 | 94.6 |
Cypress | 83.0 | 93.1 | 99.8 | 90.9 | 95.7 | 92.2 |
Korean Pine | 90.9 | 93.5 | 63.7 | 99.8 | 98.6 | 99.3 |
Korean Red Pine | 97.0 | 96.1 | 96.4 | 76.5 | 87.0 | 82.4 |
Larch | 92.6 | 94.8 | 98.3 | 81.6 | 78.4 | 80.3 |
Average | 92.5 | 95.4 | 91.6 | 89.8 | 91.9 | 89.8 |
Standard Deviation | 5.6 | 2.3 | 14.0 | 9.5 | 8.1 | 7.3 |