Leveraging Deep Learning for Identification of Medicinal Plant Species

Identifying medicinal plant species is crucial in plant taxonomy because it allows us to harness the therapeutic properties of medicinal plants for human health. This information can be beneficial for the advancement of new drugs and treatments and for promoting the sustainable use of medicinal plants. Allopathic medicine has side effects in many cases. Traditional medicine, like Ayurveda, provides alternative treatments with minor side effects and helps protect human health. The decline of many medicinal plant species highlights the need for botanical research and biodiversity conservation. Artificial intelligence builds a gap between applications of computer technology in plant taxonomy. The goal of the study is to develop a reliable deep learning system for the recognition of medicinal plant species, which can subsequently aid in their preservation and support ongoing research efforts in this field. This paper focuses on a deep learning approach based on Convolutional Neural Network (CNN) to identify three distinct categories of medicinal plants. The model proposed uses CNN architecture and two pre-trained models, InceptionV3 and ResNet50, to classify three classes of medicinal plants, namely neem, peepal, and tulsi. The CNN model has an accuracy of 97.14% and a loss of 0.1274, the ResNet50 model has an accuracy of 97.06% and a loss of 0.1474, and the InceptionV3 model has an accuracy of 100% and a loss of 0.0199 in identifying the medicinal plant species.