Enhancing Tulsi Leaf Disease with CNN and SVM-Based Classification

The research paper abstract provides an overview of the analysis's findings and results. The paper looks into disease prediction by applying a model using mint plant data. Grey Mould, Basal Roots Blight/Damping Off, Fungal Wilt and Canopy Rot, Leaves Spot, and Downy Mildew have been employed to assess the model's performance. For each class, evaluation measures like Precision, Recall, or F1-Score have been calculated, as well as support percentages and accuracy. Grey Mould (90.10%), Basal Roots Rust/Damping Off (85.48%), and Fungal Wilt and Crown Rot (84.94%) have excellent precision values, confirming this model's ability to satisfactorily categorize cases of these illnesses. The model also had high recall ratings for Leaf Spotted (85.90%) and Downy material Mildew (84.03%), demonstrating its ability to identify genuine positive occurrences. The Macro Median F1-Score of 83.15% demonstrates balanced overall class performance. The Micro Average the F1-S of 83.07% measures overall model correctness, whereas the average weighted F1-Score of 83.13% addresses class imbalances. These findings demonstrate the model's dependability and potential for predicting illness in mint vegetation, paving the path for better agricultural practices and healthier plant management.