Efficient Deep Learning Framework for Corn, Grape, and Tomato Leaf Disease Identification
Year: 2025
Type: article
Abstract: -This study introduces a robust deep learning approach for detecting plant diseases in variouscrop types through leaf image classification. The framework leverages pre-trained models and transferlearning to overcome the limitations of traditional manual inspection.A modified version of the NewPlant Diseases Dataset, comprising more than 27,000 images from 18 classes across crops like corn,tomato, and grape, was used to train and validate the models. After applying standardizedpreprocessing and offline augmentation techniques, the models are trained and evaluated usingaccuracy, precision, recall, F1-score, validation loss, and inference time. Among the tested models,MobileNetV2 delivered the highest performance with an accuracy of 98% and a macro F1 score closeto 0.98. The results confirm the robustness of the system in distinguishing healthy and diseased leaves,even across diverse categories. The proposed model is designed to be lightweight, easily deployablein the field, and suitable for real-time agricultural applicationsthat can empower farmers andagricultural stakeholders to make timely decisions, reduce crop losses, and support sustainable farmingpractices. Key Words: deep learning, transfer learning, plant disease detection, MobileNetV2, agriculture,CNNs.
Authors: Tejaswini B. Kushare, Prof. Dr. Shyamrao V. Gumaste
Institution: Savitribai Phule Pune University
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: green
OpenAlex ID: W6967713545
FWCI:
Cites:
Cited by:
Related to: