AN EFFICIENT APPROACH FOR THE CLASSIFICATION OF MEDICINAL LEAVES USING BFO AND FRVM

Herbal plants have been used for medicinal purposes since the ages. These plants also play a major role in medicines, food, perfumes and cosmetics. At present, the identification of herbal plants is purely based on the human perception of their knowledge. It may be probability of human error occurring. An efficient herb species classification system should be automatic and a convenient recognition of herbal plants which reduces the human error. The present research aims to predict the herbal plants in a very convenient and accurate way. This approach is based on the leaf shape, texture, color and its feature. Bacteria Foraging Optimization (BFO) for feature selection and Fuzzy Relevance Vector Machine (FRVM) for the classification of herbal plants are used in the proposed system. The data required for classification are computed using the MATLAB software. In the present work, ten different types of herbal leaves and twenty samples of each have been considered for the process and the classification accuracy is achieved as maximum with an efficient intelligence technique. The efficiency of the proposed method of classifying the different herbal plants gives better performance.