Classifying date fruits using the transfer learning model

• Main Authors: Mohd Adnan, Alia Nadzirah, Nasharuddin, Nurul Amelina
• Format: Article
• Language: English
• Published: Universitas Ahmad Dahlan 2024
• Online Access: http://psasir.upm.edu.my/id/eprint/113070/1/113070.pdf; http://psasir.upm.edu.my/id/eprint/113070/; https://telkomnika.uad.ac.id/index.php/TELKOMNIKA/article/view/25924
• Institution: Universiti Putra Malaysia

Date palm trees originate in many tropical regions of the world and produce dates. Each variety can be differentiated through the shape, texture, size, and colour of the fruits. People have difficulties visualising and recognising the types of date fruits because they have many varieties and species. An Android-based mobile application is being proposed to help users quickly identify the dates based on their images and expand their knowledge of dates. The date fruit species classification mobile application categorises nine different varieties of date fruits, namely Ajwa, Medjool, Rutab, Nabtat Ali, Meneifi, Galaxy, Sugaey, Shaishe, and Sokari. The classification, which is based on a transfer learning technique from a pre-trained neural network, achieved a 94.2% accuracy rate. The mobile application features a user-friendly graphical interface that makes it easy to use and understand. Users can learn about different date fruit varieties and improve knowledge retention through a mini game. The application’s usability, usefulness, and interface design were confirmed through the user acceptance survey.