CNN and Edge-Based Segmentation for the Identification of Medicinal Plants

Abstract

In Ayurveda, traditional and herbal medicine industries, it is of the utmost importance to precisely identify the appropriate medicinal plants used in the production of medicines. A medicinal plant's identity is determined by the shape, color, and texture of its leaves. SVM and KNN are existing Machine Learning (ML) models employed for this task; however, when it comes to identifying medicinal plants, they are ineffective. For medicinal plant identification, complex properties like leaf form, color, and texture can be difficult to accurately represent using traditional methods like KNN and SVM. In image recognition tasks, large datasets are common, and some approaches may not scale well under those circumstances. They might become prohibitively expensive to compute, rendering them useless for everyday use. The development of strong characteristics for plant recognition is frequently a time-consuming process, despite the fact that traditional models heavily rely on feature engineering. Hence, a promising approach to quickly identify the beneficial plant species is presented in this study. A Convolutional Neural Network (CNN) is the solution proposed here to identify the medicinal plants. CNNs are appropriate for processing largescale data, for example, leaf structure, variety, and surface discovery, without requiring any prerequisite for broad element design due to their capacity to gain and concentrate requested qualities from the input images. CNNs are more adaptable to the real world because they can be trained effectively on large datasets and can be parallelized. When contrasted with different models, the one developed in this research study improves at characterizing the restorative plant.