Journal of Ayurveda and Integrative Medicine · 2024 · 11 citations · 34 references
ComputingBotanyComputer ScienceComputer VisionDrug DiscoveryMachine LearningDeep LearningMedicinal PlantsPlant IdentificationTransfer LearningDeep Learning-based ClassificationDeep Learning-based IdentificationIndigenous Medicinal Plant KnowledgeMedicinal Plant ClassificationMedicinal Plant ConservationMedicinal Plant TaxonomyTraditional Healers' Knowledge
BACKGROUND: Medicinal plants are crucial for traditional healers in preparing remedies and also hold significant importance for the modern pharmaceutical industry, facilitating drug discovery processes. Accurate and effective identification and classification of Ethiopian indigenous medicinal plants are vital for their conservation and preservation. However, the existing identification and classification process is time-consuming, and tedious, and demands the expertise of specialists. Botanists traditionally rely on traditional and experience-based methods for identifying various medicinal plant species. OBJECTIVE: This research aims to develop an efficient deep learning model through transfer learning for the identification and classification of Ethiopian indigenous medicinal plant species. MATERIALS AND METHODS: A custom dataset of 1853 leaf images from 35 species was prepared and labeled by botanist experts. Experiments have been done with the use of pretrained deep learning models, specifically VGG16, VGG19, Inception-V3, and Xception. RESULTS: The results demonstrate that fine-tuning the models significantly improves training and test accuracy, indicating the potential of deep learning in this domain. VGG19 outperforms other models with a test accuracy of 94%, followed by VGG16, Inception-V3, and Xception with test accuracies of 92%, 91%, and 87%, respectively. The study successfully addresses the challenges in the identification and classification of Ethiopian indigenous medicinal plant species. CONCLUSION: With an inspiring accuracy performance of 95%, it can be concluded that fine-tuning emerged as a highly effective strategy for boosting the performance of deep learning models.
34
Going deeper with convolutions
Christian Szegedy, Wei Liu, Yangqing Jia et al. · 2015 · 46.8K citations
ImageNet Large Scale Visual Recognition Challenge
Olga Russakovsky, Jia Deng, Hao Su et al. · International Journal of Computer Vision · 2015 · 40.2K citations
Xception: Deep Learning with Depthwise Separable Convolutions
François Chollet · 2017 · 19K citations
Revisiting Unreasonable Effectiveness of Data in Deep Learning Era
Chen Sun, Abhinav Shrivastava, Saurabh Singh et al. · 2017 · 2.3K citations