Title: A method for identification and classification of medicinal plant images based on level set segmentation and SVM classification
Authors: Suvarna S. Nandyal; Basavaraj S. Anami; A. Govardhan; P.S. Hiremath
Addresses: Department of CSE, PDA College of Engineering, Aiwan-E-Shahi Area, Gulbarga-585102, Karnataka, India. ' K.L.E. Institute of Technology, Hubli-580030, Karnataka, India. ' JNTU College of Engineering, Kukatpally, Hyderabad-500 085, Andrapradesh, India. ' Department of P.G. Studies and Research in Computer Science, Gulbarga University, Gulbarga-585106, Karnataka, India
Abstract: This paper presents a methodology for identification and classification of images of the medicinal plants based on level set segmentation. The medicinal plants are identified using structural features, namely, height, shape, size of leafy part, flowers, fruits, and branching patterns. In this work, the level sets are used for segmentation of images of medicinal plants. The two segments, namely, leafy part (canopy) and stem, are obtained. The geometrical ratios of length to width of leafy and stem parts of images are used as features. The classification of images of medicinal plants into herbs, shrubs and trees using minimum distance, neural network and SVM classifiers is performed. The experiments are carried on 400 images of medicinal plants of different classes, such as Calotropis gigantea, Aloe vera, Catharantus roseus, Carica Papaya, Azadirachita indica and Cocos nucifera. The classification accuracies obtained by different classifiers are compared. It is observed that the combination of level set segmentation and SVM classifier yielded better classification results. The knowledge of these medicinal plants is useful for practitioners of Ayurveda system of medicine, botanists and common man for home remedies.
Keywords: medicinal plants; herbs; shrubs; trees; geometric features; level set segmentation; minimum distance classifier; plant classification; plant images; SVM classification; support vector machines; image segmentation.
International Journal of Computational Vision and Robotics, 2012 Vol.3 No.1/2, pp.96 - 114
Published online: 07 Apr 2012 *Full-text access for editors Access for subscribers Purchase this article Comment on this article