Table 1 Comparison of EfficientNet, ResNet, and DenseNet.

From: Adaptive hybrid segmentation combined with meta heuristic optimization in transfer learning for plant leaf disease classification

Model

Advantages

EfficientNet (efficient network)

It is appropriate for real-time applications as it makes use of compound scaling to maximize accuracy and efficiency

ResNet (residual network)

It enhances feature extraction during deep network training by using residual learning to counteract disappearing gradients

DenseNet (densely connected convolutional network)

It improves classification accuracy and illness pattern identification by strengthening feature reuse through thick layer connections

Back to article page