Table 2 Comparative analysis of the proposed work and existing studies on plant disease detection
Application Field & Problem | Problem & Reference | Dataset | Models Used for Classification | Accuracy (%) | XAI Methods Used | Result Interpretation Method | Remarks on Result Interpretation Methods Followed |
|---|---|---|---|---|---|---|---|
Agriculture - Plant Disease Detection | Fruit leaf disease classification47 | PlantVillage73 | ResNet | 99.89 | GradCAM, SmoothGrad, LIME | – | Results are interpreted manually. Quantitative measures are not provided. |
Tomato leaf disease detection45 | Tomato leaf dataset74 | EfficientNetB5 | 99.07 | GradCAM | – | ||
Mulberry leaf disease detection46 | BSDB | PDS-CNN | 95.05 | SHAP | – | ||
Potato leaf disease detection75 | Farms in West Bengal | CNN-SVM, DenseNet169 | 99.98 | LIME, SHAP | Visual interpretation | ||
Sunflower leaf disease detection76 | Sunflower dataset77 | VGG19 + CNN | 93.00 | LIME | – | ||
Crop disease classification78 | Thermal image dataset | PlantDXAI | 98.55 | CAM | – | ||
Plant leaf disease detection57 | PlantVillage79 | ResNet-50 | 99.99 | GradCAM | – | ||
Tomato fruit quality classification56 | Tomatoes dataset80 | MobileNetV2 | 98.00 | GradCAM | Quantitative analysis | Authors did not explain formula for Match Ratio | |
Proposed Work | – | Rice leaf disease dataset | ResNet50, Xception, InceptionResNetV2, DenseNet201, AlexNet, VGG16, InceptionV3, EfficientNetB0 | 99.13 | LIME | Quantitative analysis | IoU, Jaccard Distance, Dice Similarity, Precision, Sensitivity, Specificity, MCC |
