CASSAVA DISEASE DETECTION USING CONVOLUTIONAL NEURAL NETWORK.

Abstract Cassava, a staple food for millions, faces significant yield threats due to devastating leaf diseases such as Cassava Mosaic Disease, Brown Streak Disease, and Bacterial Blight. Traditional detection methods are labor-intensive, subjective, and slow, limiting timely intervention. To address these challenges, this study proposes a Convolutional Neural Network (CNN)-based cassava disease detection model capable of automatically classifying leaf images into five categories: Cassava Bacterial Blight, Brown Streak Disease, Green Mottle, Mosaic Disease, and Healthy. A labeled dataset sourced from Kaggle was preprocessed with image resizing, normalization, and data augmentation to improve generalization. The dataset was then split into training, validation, and testing sets. A CNN model, enhanced through transfer learning with MobileNetV2, was trained using the categorical cross-entropy loss function and the Adam optimizer. Hyperparameters such as learning rate, batch size, and epochs were carefully tuned. The model was evaluated using accuracy, precision, recall, F1-score, and confusion matrix, with k-fold cross-validation providing robustness. Findings showed the model achieved a training accuracy of 98.3%, validation accuracy of 96.5%, and test accuracy of 95.8%. The confusion matrix revealed minimal misclassifications, with the highest errors occurring between visually similar diseases such as Green Mottle and Mosaic Disease. Overall precision, recall, and F1-scores exceeded 94%, confirming the model’s excellent performance.The CNN-based system demonstrates superior accuracy and scalability compared to traditional methods and baseline models. It offers a powerful, reliable, and automated solution for early cassava disease detection, supporting timely intervention and improved yield outcomes. Deployment on lightweight platforms, combined with further data expansion and explainable AI methods, will enhance its future usability for farmers and agricultural experts in promoting food security.