<![CDATA[This study aims to develop and evaluate a convolutional neural network (CNN)-based model for classifying corn leaf diseases using a simple yet effective architecture. Four disease classes were considered: healthy, gray leaf spot, leaf blight, and common rust. A dataset comprising 13,136 images was obtained from the open-source PlantVillage Dataset and processed using class balancing techniques to mitigate prediction bias. Each image was resized to 256×256 pixels, normalized, and split into training (80%) and testing (20%) sets. The proposed CNN architecture consists of four convolutional layers with progressively increasing filters (16, 32, 64, 128), followed by max pooling, dropout, and two fully connected layers. The model was trained for 50 epochs using the Adam optimizer with categorical cross-entropy as the loss function. Performance evaluation, based on accuracy, precision, recall, and F1-score, achieved an accuracy of 97.18% with consistently high metrics across all classes. The results were further visualized using a confusion matrix and classification report. Finally, the trained model was deployed in a Flask-based web application, enabling users to upload corn leaf images and receive automated detection results. These findings demonstrate that a simple CNN architecture can achieve high accuracy in classifying corn leaf diseases and holds significant potential for integration into digital plant disease monitoring systems.]]>
Copyrights © 2025
