Title: Deep Learning-Based Classification Of Grape Leaf Diseases Using Xception: A Precision Agriculture Approach
Authors: Mohammed S. Abunasser and Samy S. Abu-Naser
Volume: 10
Issue: 6
Pages: 1-9
Publication Date: 2026/06/28
Abstract:
Artificial intelligence-driven plant disease detection has emerged as a critical tool for improving agricultural productivity and ensuring food security. This study proposes an enhanced deep learning framework based on a fine-tuned Xception architecture for the automated classification of grape leaf diseases. A balanced dataset of 8,000 images across four classes-Black Rot, ESCA, Leaf Blight, and Healthy leaves, was utilized to ensure unbiased learning. The dataset was partitioned into training (60%), validation (20%), and testing (20%) subsets. To improve generalization and robustness, the proposed model integrates transfer learning, data augmentation, batch normalization, dropout regularization, and L1-L2 penalties. In addition to conventional evaluation metrics (accuracy, precision, recall, and F1-score), the model performance was further validated using confusion matrix analysis and Receiver Operating Characteristic (ROC) curves. Experimental results demonstrate that the proposed framework achieves an accuracy of 99.90% and an F1-score of 99.85% on the test set. Despite the high performance, additional validation highlights potential limitations related to dataset complexity and real-world variability. Comparative analysis with existing models confirms the effectiveness of the proposed approach. The findings suggest that the model can be effectively deployed in precision agriculture systems, including mobile and edge-based applications, for real-time disease monitoring.