<![CDATA[Purpose: This research aims to develop an artificial intelligence (AI) model based on the Convolutional Neural Network (CNN) to classify Uvaria plant species, a tropical genus native to Indonesia. The study addresses the challenge of limited datasets for automatic classification in tropical plant identification. Methodology/approach: Images of Uvaria plants were collected directly from their natural habitat and categorized into four primary classes: leaves, stems, twigs, and trees. The dataset comprises 400 labeled images, split into training (279 images, 70%), validation (40 images, 10%), and testing (81 images, 20%). The CNN model was trained for 200 epochs, using data preprocessing techniques such as normalization and augmentation to improve performance. Results/findings: The CNN model achieved an accuracy of 90% on the test set, indicating strong performance in classifying the four categories of Uvaria plant components. The model showed particularly consistent results in distinguishing between leaves and twigs. Conclusion: Despite the relatively small dataset, the results demonstrate that the CNN algorithm is capable of accurately classifying images of Uvaria species. The dataset is considered sufficient to build an effective classification model. Limitations: The main limitation of this study is the limited number of images, which may restrict the model’s ability to generalize to broader or more varied data in real-world conditions. Contribution: This research contributes to the development of AI-based tools for identifying tropical plant species. It offers a practical model and dataset that can support biodiversity monitoring, environmental research, and conservation efforts in Indonesia and similar tropical regions.]]>
