Region-Based Fully Convolutional Networks for Medical Plants Identification

In India, the cultivation and harvesting of medicinal herbs has a long history. Many beneficial medicinal herbs are mostly found in the Indian Forest. Because of their vital function in sustaining human existence, medicinal plants have long been the focus of much study and analysis. The process of accurately identifying medicinal plants is tedious and requires a great deal of skill. As a result, a vision-based method could help researchers and laypeople identify herb plants quickly and accurately. This study proposes a deep learning-based approach to classify medicinal plant species using R-FCN. The model leverages two primary components: the FCN for feature extraction and the RPN for region selection. The suggested method processes plant images by detecting significant regions and classifying them into corresponding species categories. The combination of object detection and classification improves the accuracy of species identification by focusing on key plant structures. Medicinal leaves are entirely automatically identified using the R-FCN based deep learning method. The experimental findings illustrate the effectiveness of the R-FCN model in identifying medicinal plants with high precision, making it a promising solution for automated plant recognition systems.This led to a $\mathbf{9 9. 8 2 \%}$ accuracy rate in the DL model's ability to distinguish between medicinal leaves.