<![CDATA[This study focuses on developing an Android-based recommender system using convolutional neural networks (CNNs) to select high-quality grapes. The main objective of this study is to compare the performance of two popular CNN architectures, VGG16 and ResNet18, in classifying the quality of sour grapes. The subjective and time-consuming nature of conventional methods prompted us to search for a more efficient solution.The dataset used consists of 282 images of green grapes. The evaluation results show that the VGG16 model achieves 93% accuracy in classifying grape quality, outperforming the ResNet18 model with only 82% accuracy. These results indicate that the VGG16 architecture is more suitable for this classification task. The development of this system is expected to contribute to smart agricultural automation to improve efficiency and support the food industry.]]>
Copyrights © 2025
