International Journal of Computer Applications (0975 - 8887) Volume 62 - No. 12, January 2013 Texture Feature Extraction for Identification of Medicinal Plants and Comparison of Different Classifiers C. H. Arun Department of Computer Science NMC College Marthandam, India W. R. Sam Emmanuel Department of Computer Science NMC College Marthandam, India D. Christopher Durairaj Department of Computer Science VHNSN College Virudhunagar, India ABSTRACT This paper presents an automated system for recognizing the medicinal plant leaves that are taken from the suburbs of the western ghats region. The dataset comprises of 250 different leaf images, of five species. Texture analyses of the leaf images have been done in this work using the feature computation. The features include grey textures, grey tone spatial dependency matrices(GTSDM) and Local Binary Pattern(LBP) operators. For each leaf image, a feature vector is generated from the statistical values. 70% of the images in the dataset are the training dataset and the rest are included in the test set. Six different classifiers are used to classify the plant leaves based on feature values. When features are combined without any preprocessing, it yielded a classification performance of 94.7%. General Terms: image processing, pattern recognition Keywords: image classification, texture features, plant identification 1. INTRODUCTION Herbal medicines are gaining popularity worldwide as they are safe to human health and affordable. World Health Organization estimates that 80% of people in Asia and Africa rely on herbal medicines for some aspects of their primary health care [1]. There are diverse medicinal plant varieties in lush forests of southern india which are used as medicines to a number of sicknesses like common cold, allergies etc. Precise identification of the respective plant leaf is vital in treating the patients. The present work explores the use of texture features to identify the appropriate medicinal leaves automatically. Textures are a pattern of non-uniform spatial distribution of differing image intensities [11], which focus mainly on the individual pixels that make up an image. Texture is defined by quantifying the spatial relationship between materials in an image. Texture analyses is done by using various approaches like statistical, fractal, and structural. Statistical type includes techniques like grey-level histogram, grey-level co-occurence matrix [20], local binary pattern operator, auto-correlation features and power spectrum. Grey texture features or the first order statistical features are used in image classification [15]. Haralick et al [14], described texture features based on grey-level co-occurence matrix(GLCM) and are used in various applications including Land cover classification [22]. Ojala et al [19] used texture feature based on local binary pattern(LBP). Textures are useful in identifying plant leaves without the need for an expert. [26][21][12]. Plant leaves are recognized based on the features of leaf images[7]. Classification of leaf epidermis microphotographs is done using texture features [23]. Basavaraj et al [3], Sandeep et al [24] identified medicinal plant leaves using textures. Classification approaches like Stochastic Gradient Descent(SGD) [25][18] , kNearest Neighbour(kNN) [4][2], Support Vector Machines(SVM) [10], Decision Trees(DT) [5], Extra Trees(ET) [13], Random Forests(RF)[8][6] can be used as Image classifiers. The objective of the present work is to compute texture features based on grey-level, the Grey Tone Spatial Dependency Matrix(GTSDM) and the Local Binary Pattern(LBP) operators and to obtain the best combination of these, for automatically identifying the right medicinal plant. Five medicinal plant leaves are considered and few of their medicinal uses are given below. A sample of the medicinal plant leaf dataset from different plant leaves are given in Fig. 1. (0) Desmodium gyrans - Antidote for snake poison, effective for heart diseases, rhematic complaints, Diabetes and skin ailments. (1) Butea monosperma - promotes diuresis and anthelmintic, treats leucorrhea and diabetes. (2) Malpighia glabra - help lower blood sugar, increases collagen and elastin production, treats diarrhea, dysentery, and liver problems. (3) Helicteres isora - help treat intestinal complaints, colic pains and flatulence. (4) Gymnema sylvestre - suppresses the sensation of sweet, anti-diabetic. The rest of the paper is organized as follows: Methodology of the work is discussed in section 2. This section includes the details of texture based image classification and the description of the dataset. Design and implementation of the work are discussed in section 3. The results and discussion are given in section 4, which is followed by the conclusion. 1