Image-Based Cashew Leaf Disease Detection Using an Optimized DenseNet Framework
Isaac Angelo M. Dioses, Wendell M. Castillo, Jesusimo L. Dioses, Jolan Sy, Desiray Nayga, Carmelo Alejo Bisquera, Darios B. Alado
Abstract
This study explores the application of deep learning for the detection and classification of cashew leaf diseases using the DenseNet169 convolutional neural network architecture. A total of 6,614 cashew leaf images were gathered from an open-source dataset comprising five categories: Anthracnose, Gummosis, Leaf Miner, Red Rust, and Healthy. The dataset was divided into 70 % for training, 15% for validation, and 15% for testing. Data augmentation techniques were applied to enhance model generalization and reduce overfitting. The model’s performance was assessed using precision, recall, F1-score, and a confusion matrix. Experimental results show that DenseNet169 achieved an overall classification accuracy of 95.06 %, demonstrating consistent and reliable performance across all disease classes. The study confirms that DenseNet169 is an effective model for classifying cashew leaf diseases and provides a strong basis for the development of automated disease detection systems that can assist in improving cashew production and crop management in the Philippines.