Fig 1.
Workflow framework for diagnosing MDM and providing field advisory.
Fig 2.
Architecture of the DSS employing VGG16 for MDM identification.
Table 1.
Comparative performance of thirteen machine-learning and deep-learning models for maize downy mildew detection.
Fig 3.
Comparative accuracy and precision of the top 10 models for MDM recognition.
Fig 4.
Evaluation of recall, F1-score, and AUC–ROC metrics for MDM detection models.
Fig 5.
Algorithmic comparison for feature extraction from MDM image datasets.
Fig 6.
Error rate and variance analysis of VGG16 model performance in MDM detection.
Fig 7.
Confusion matrix representation of VGG16 outcomes for MDM identification.
Fig 8.
Learning curve trends of VGG16 during MDM detection.
Fig 9.
t-SNE visualization of extracted features for MDM classification.
Fig 10.
Grad-CAM–driven interpretability analysis of VGG16 in MDM detection.
Table 2.
Reducing Avoidable Yield Loss in Maize through Fungicide-Based Advisory Measures Against MDM Disease.
Fig 11.
Yield Advantage and Benefit-Cost Ratio Under Fungicide Advisory for MDM Management.