Featured based Segmentation of Color Textured Images using GLCM and Markov Random Field Model

In this paper, we propose a new image segmentation World Academy of Science, Engineering and Technology International Journal of Computer, Electrical, Automation, Control and Information Engineering Vol:5, No:5, 2011 427 International Scholarly and Scientific Research & Innovation 5(5) 2011 scholar.waset.org/1999.4/6017 In te rn at io na l S ci en ce I nd ex , C om pu te r an d In fo rm at io n E ng in ee ri ng V ol :5 , N o: 5, 2 01 1 w as et .o rg /P ub lic at io n/ 60 17 interpreted in transformed spaces like HSV, YIQ, Ohta (I1, I2, I3), CIE (XYZ, La*b*), these have been utilized in image segmentation. In this work, we propose a new approach, which adopts the features of both gray level co-occurrence matrix and Markov Random field model using Ohta color space to segment color textured image. GLCM represents the distance and angular spatial relationships over an image sub-region of specified size from which several textural measures may be computed. These measures are considered in classifying the textured image. In MRF based segmentation, the most popular criterion for optimality has been the maximizing a posteriori probability (MAP) distribution criterion. Simulated annealing (SA) and iterated conditional modes (ICM) algorithm are two unremarkably used methods for pixel labeling, among the existing MAP criterion algorithms. SA can converge to global optimum, but suffers from intensive computation. On the other hand, ICM has the ability of faster convergence but the results obtained heavily depend on initial state. Hence in our proposed work GLCM feature matrix obtained in Ohta colour space provides a very good initialization for ICM algorithm. Section II reports GLCM and texture measures. MRF model based segmentation is explained in section III. Section IV describes the proposed segmentation approach based on GLCM and MRF. Results are presented in section V and conclusion is presented in section VI. II. GREY LEVEL CO-OCCURRENCE MATRIX AND TEXTURAL MEASURES Texture features based on GLCM are an efficient means to study the texture of an image. Given the image composed of pixels each with an intensity, the GLCM is an illustration of how frequently different combinations of grey levels concur in an image. A GLCM denote the second order conditional joint probability densities of each of the pixels, which is the probability of occurrence of grey level i and grey level j within second order statistics are calculated for all pair wise combinations of grey levels.. Generated the GLCM, 14 types of texture features have been defined by Haralick et. al. [3]. The depiction of the texture information is then extracted by these series of texture statistics computed from GLCM. In our study we have looked at eight conventional measures. These are described as follows.