Comparative Performance of CNN From Scratch and Fine-Tuned MobileNetV2 for Tomato Leaf Disease Classification

This dataset contains tomato leaf images from the PlantVillage dataset,organized into 10 classes for multiclass image classification. This version has been prepared through an image validation, class-namenormalization, SHA-256 duplicate detection, consistent file renaming,and stratified data splitting process. The dataset is divided into approximately: - 80% training data- 10% validation data- 10% testing data The dataset contains the following classes: 1. bacterial_spot2. early_blight3. healthy4. late_blight5. leaf_mold6. septoria_leaf_spot7. target_spot8. tomato_mosaic_virus9. tomato_yellow_leaf_curl_virus10. two_spotted_spider_mite The accompanying reports provide information about class distribution,corrupt files, duplicate files, dataset splitting, and class-name mapping. The original images originate from the PlantVillage dataset. This recordrepresents a cleaned, validated, deduplicated, renamed, and pre-splitversion prepared for reproducible machine learning and computer visionexperiments.