Table 1 Comparison of the proposed work with various deep learning approaches for rice leaf disease detection.

CNN based Approach

^{7,8,9,10,11,12,60,61,62,63,64,65}

Precision, Recall, F1-score, specificity, Accuracy, Average Precision (mAP), Area of Precision (AP), Mean Square Error (MSE), Loss function and Root Mean Square Error

Not used

XAI techniques are not used to visualize and understand the decision-making process.

Transfer Learning Approach

^{13,14,15,16,17,18,19,20,21,22,23,24,66,67,68}

Area Under Curve (AUC), Precision, Recall, Accuracy, F1-Score, Specificity, Matthews Correlation Coefficient (MCC), Falser Positive Rate (FPR), Negative Predictive Value

Not used

XAI techniques are not used to visualize and understand the decision-making process.

⁵³

Precision, Recall, Accuracy, F1-Score

Intermediate Class Activation Map (ICAM)

XAI techniques are used for visual explanations, but no quantitative metrics are used to compare these visual explanations.

⁵¹

Precision, Recall, Accuracy, F1-Score

GradCAM

Ensemble Learning Approach

^26,27

Precision, Recall, Accuracy, F1-Score, Specificity, Support

Not used

XAI techniques are not used to visualize and understand the decision-making process.

⁵⁴

Precision, Matthews Correlation Coefficient (MCC), Recall rate, Accuracy, F1-score,

GradCAM, GradCAM++, Guided Backpropagation

XAI techniques are used for visual explanations, but no quantitative metrics are used to compare these visual explanations.

⁵²

Accuracy, Precision, F1-score, Specificity

GradCAM++, Score-CAM

Hybrid Approach

^28,29,30,69

AUC, Precision, Recall, F1-Score, Specificity, Kappa coefficient, Accuracy

Not used

XAI techniques are not used to visualize and understand the decision-making process.

Proposed Work

–

Accuracy, Precision, Recall, F1-score, Specificity

LIME

XAI techniques are used for visual explanations. Results are compared with both qualitative and quantitative measures