Abstract—This paper presents X-CoffeeNet, a novel deep learning framework for classifying coffee bean images by integrating MobileNetV2 and Explainable Artificial Intelligence (XAI) techniques. The model focuses on accurately classifying Arabica, Robusta, and Liberica beans while ensuring transparency in its decision-making process. The main novelty lies in combining a lightweight convolutional neural network with Local Interpretable Model-agnostic Explanations (LIME), allowing users to visualize and understand which parts of the image influence predictions. To improve generalization and reduce overfitting, the model uses image augmentation methods such as rotation, flipping, zooming, and brightness adjustment. MobileNetV2 is selected for its efficiency and low computational cost, making X-CoffeeNet suitable for deployment on mobile or embedded systems. Experimental evaluation on a dataset of 3,000 images shows that X-CoffeeNet achieves 100% accuracy, precision, recall, and F1-Score across all classes, highlighting its reliability and performance. Beyond coffee classification, the framework has broader applicability in agriculture and food industry automation where both accuracy and interpretability are essential.
 
Keywords—coffee bean classification, deep learning, Explainable Artificial Intelligence (XAI), MobileNet, image processing, computer vision


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).