Author | Year | Model | Limitation |
---|---|---|---|
[18] | 2024 | PDDNet-AE, PDDNet-LVE | No data of handling overfitting |
[19] | 2023 | AgriDet framework with INC-VGGN | Lack of comparison with newer models |
[20] | 2023 | DenseNet and EfficientNet | Minor drop in testing accuracy compared to training accuracy |
[21] | 2022 | DenseNet-121, ResNet-50, VGG-16, Inception V4 | No data of solving overfitting or generalization |
[22] | 2022 | Deep CNN | Focused only on semantic segmentation, not end-to-end detection |
[23] | 2024 | Visual Transformer architectures | Limited generalization outside specific datasets |
[24] | 2024 | CNN with six layers and Dense ANN | No information of comparison with modern transformer models |
[25] | 2024 | LeafyGAN | Lower performance on PlantDoc dataset compared to PlantVillage |
[26] | 2024 | Transformer block model with inception architecture | Requires high computational resources due to transformer architecture |
[27] | 2024 | CBAM + MAE | No indication of real-time efficiency or generalizability to other datasets |