<![CDATA[High postharvest food loss in horticultural commodities, particularly bananas, mangoes, and papayas, remains a significant challenge in postharvest handling processes. One of the main causes is the determination of fruit ripeness levels, which is still commonly performed manually and is therefore prone to subjectivity and inconsistency. This study aims to implement a Vision Transformer (ViT) architecture for fruit ripeness classification and to support an automated quality control system through an Android-based application. The research involved dataset collection using smartphone cameras, image preprocessing, data augmentation, Vision Transformer model training, and performance evaluation using confusion matrix, accuracy, precision, recall, F1-score, robustness testing on non-target images, and inference latency measurement. The trained model was then integrated into a mobile application for automatic ripeness prediction. The results show that the Vision Transformer model performs well in fruit ripeness classification. The model achieved an average test accuracy of 95.31% and an overall accuracy of 94.94%, with a precision of 0.94, recall of 0.96, and F1-score of 0.95. In addition, the model successfully identified non-target images with precision, recall, and F1-score values of 1.00 and achieved an average inference latency of 0.74 seconds. These findings indicate that Vision Transformer is effective for fruit ripeness classification and has strong potential to support faster, more consistent, and objective postharvest fruit quality control.]]>
Copyrights © 2026
