Optimizing Species Recognition in Medicinal Plants: A Comprehensive Evaluation of Deep Learning Models

Medicinal plants have the potential to both significantly progress the pharmaceutical and medical industries and shield people from fatal diseases like cancer and cardiovascular conditions. Since there are hundreds of kinds of plants in the globe and because their characteristics: -like color, texture, and form are all the same, it might be difficult to identify them. The manual process is laborious and time-consuming. In view of current technological breakthroughs, more efficient methods for meeting the need for species identification have required to be devised. Pattern recognition and digital image processing are two examples of these methods. Despite several prior research, there are still difficulties in reliably automating plant species identification in a sequential and adaptive manner to identify a variety of plant species. In this article, the most recent deep learning models (VGG19, InceptionV3, LeNet, and Resnet50) are compared for medicinal plant identification and categorization. The system is trained using 1835 leaf samples from 30 distinct medicinal plants. For every model, confusion and performance metrics were used. In contrast to the other models in this investigation VGG19 fared better, with an accuracy rate of 99.85%.