Indonesia Medicinal Plant Recognition using Deep Learning: A Computational Study

Medicinal plants are beneficial plants that can serve as alternatives to traditional medical treatments for various illnesses. Deep learning models have emerged as powerful tools in the field of medical research, offering innovative approaches for the identification and classification of various entities, including medicinal plants. The work clarifies the effectiveness of deep neural networks in computer vision, specifically in relation to the identification of images of medicinal plants. Our paper focuses on identifying medicinal plants using deep-learning models such as InceptionV3 and VGG19. The study proposes the comparative performance of the different models such as ResNet34, DenseNet121, and VGG11. The previous work models, DenseNet121 achieves the highest accuracy of 87.4% on the Indonesia Medicinal Plant Dataset, followed by ResNet34 with 85.65%, and VGG11-bn with 82%. The evaluation results of our work, based on the Indonesian medicinal plant's dataset, highlight the effectiveness of pre-trained models, showcasing VGG19 accuracy of 89.65 % and Inception-V3's remarkable accuracy of 98.86 %.