<![CDATA[This study applies a transfer-learning approach using the VGG16 architecture to classify three Papuan endemic orchid species—Dendrobium spectabile, Dendrobium lineale, and Dendrobium mirbelianum. A total of 810 field-photographed images were collected, followed by preprocessing and data augmentation to enhance data diversity. The VGG16 model pretrained on ImageNet was used as a fixed feature extractor by freezing its convolutional layers and removing the fully connected layers, while a custom classification head was added to distinguish among the three species. Experimental results demonstrated a validation accuracy of 94.44% and a macro-average F1-score of 0.94, confirming the robustness of the model under limited-data conditions. These findings suggest that transfer learning using VGG16 can effectively support orchid species recognition and serve as a foundation for developing AI-based biodiversity monitoring and conservation systems in Indonesia]]>
Copyrights © 2025
