Table 1 Result of validation and test using five pre-trained CNN models.
From: Construction of deep learning-based disease detection model in plants
Step | Crop | Pre-trained model | Accuracy | ||
|---|---|---|---|---|---|
Validation | Test | ||||
I | Crop classification | All | ResNet50 | 92.34% | 91.84% |
AlexNet | 97.62% | 96.87% | |||
GoogleNet | 98.16% | 99.08% | |||
VGG19 | 96.86% | 98.71% | |||
EfficientNet | 98.54% | 99.33% | |||
II | Disease detection | Bell Pepper | ResNet50 | 100.00% | 98.32% |
AlexNet | 98.96% | 99.16% | |||
GoogleNet | 100.00% | 100.00% | |||
VGG19 | 99.74% | 99.58% | |||
EfficientNet | 99.48% | 99.58% | |||
Potato | ResNet50 | 100.00% | 99.45% | ||
AlexNet | 99.31% | 98.90% | |||
GoogleNet | 99.66% | 99.45% | |||
VGG19 | 100.00% | 100.00% | |||
EfficientNet | 100.00% | 99.45% | |||
Tomato | ResNet50 | 99.68% | 99.75% | ||
AlexNet | 99.53% | 99.45% | |||
GoogleNet | 99.62% | 99.62% | |||
VGG19 | 99.12% | 99.62% | |||
EfficientNet | 99.62% | 98.23% | |||
III | Disease classification | Potato | ResNet50 | 99.25% | 98.80% |
AlexNet | 98.88% | 99.40% | |||
GoogleNet | 99.63% | 99.40% | |||
VGG19 | 99.63% | 99.40% | |||
EfficientNet | 97.75% | 99.40% | |||
Tomato | ResNet50 | 92.92% | 87.80% | ||
AlexNet | 92.47% | 95.45% | |||
GoogleNet | 94.29% | 95.81% | |||
VGG19 | 95.66% | 95.08% | |||
EfficientNet | 96.35% | 97.09% | |||
