Medicinal Plant Identification by Leaf Structure Using Ensemble Methods on Deep Learning Algorithms

Abstract

India’s rich biodiversity includes a vast range of medicinal plants, widely used in traditional herbal medicine. However, the identification of these plants remains challenging. This study focuses on identifying medicinal plants using their leaf structure through deep learning and ensemble methods. Various pre-trained models, including VGG-16, VGG-19, ResNet50, InceptionV3, and EfficientNetB2, were evaluated with optimization techniques like Adam, SGD, and AdaDelta. The Adam optimizer delivered superior performance, achieving an accuracy of 97%. The research highlights the advantages of automated medicinal plant identification, providing significant potential for healthcare and biodiversity preservation.