<![CDATA[This research develops an image-based tomato ripeness classification system using Convolutional Neural Network (CNN). The dataset consists of 3600 tomato images classified into three classes, namely unripe, half-ripe, and ripe. These images were obtained from publicly available datasets, specifically 2,550 images from Kaggle and 1,050 images from GitHub. Before testing, a pre-processing stage is carried out which includes resizing to reduce the size of the image, cropping to focus on the main object, namely tomatoes, and augmentation to increase the generalization ability of the model. The CNN model is built with an architecture that is capable of extracting visual features automatically through the learning process. The training results for 10 epochs show that the validation accuracy reaches 99.39%, with a loss of 4.16%, indicating that the model is able to learn well without significant overfitting. These results show that CNNs perform well in accurately identifying tomato maturity, while emphasizing the possibility of further improving the model through transfer learning and optimization to handle more complex data]]>
Copyrights © 2025
