Texture dominant approach for identifying ayurveda herbal species using flowers

Abstract

Ayurveda treatments are used by people all over the world for centuries. These treatments are made by using parts of medicinal plants or herbal species. Therefore, it is important to know some vital Ayurveda plants for our day today life for simple herbal treatments. This paper presents a methodology to identify herbal species using their flower images. There are studies have been done to identify plants that most of them are done by using leaves and those are not specifically for Ayurveda plants. The approach extracts color, shape and different texture features from flower images and creates three different feature vectors with Haralick, Tamura and Gabor textures. Then classify each with different classifiers such as SVM, Decision Trees and K-NN. The results are compared with each other’s individual performances in order to use them for identifying Ayurveda plants.