Abstract—Nutmeg seed quality plays a key role in determining the economic value and export potential of Indonesian spice products. Manual assessment of quality is often subjective, time-consuming, and depends on the evaluator’s experience. This study introduces an explainable deep learning method designed to automatically and transparently evaluate nutmeg image quality. The model employs the EfficientNet-B0 architecture trained on 300 nutmeg images categorized into three quality levels: good, moderate, and rotten. To enhance interpretability, Explainable Artificial Intelligence (XAI) techniques—specifically Gradient-Weighted Class Activation Mapping (Grad-CAM) and local Interpretable Model-Agnostic Explanations (LIME)—were applied to highlight the image regions most influential in the model’s decisions. Model performance was evaluated using accuracy, precision, recall, F1-score, and mean Average Precision (mAP). The EfficientNet-B0 model, coupled with XAI methods, achieved an overall accuracy of 78% and a mAP of 70.29%, with Grad-CAM and LIME visualizations consistently highlighting the key visual features that determine nutmeg quality. These findings demonstrate that integrating deep learning with XAI can produce an objective, efficient, and dependable quality assessment system, which offers potential applications to other agricultural products.
 
Keywords—explainable deep learning, EfficientNet-B0, Gradient-Weighted Class Activation Mapping (Grad-CAM), Interpretable Model-Agnostic Explanations (LIME), nutmeg seed quality, artificial intelligence, smart agriculture


Copyright © 2026 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).