WSEAS Transactions on Biology and Biomedicine
Print ISSN: 1109-9518, E-ISSN: 2224-2902
Volume 11, 2014
On Application of Entropy Characteristics to Texture Analysis
Authors: ,
Abstract: A method of texture analysis based on using the Kullback-Leibler divergence is discussed. A digital image is described by a discrete probability distribution. We consider a group of direct multifractal transforms relating to the distribution and for two given images calculate a vector of Kullback-Leibler divergences between the initial distributions and their direct multifractal transforms. The method is illustrated on the example of the Serpinsky carpet. Numerical experiments were performed for the Brodatz textures and for two classes of biomedical preparation images — healthy kidney and kidney with chronic pyelonephritis. In each class divergence vectors between pairs of images are calculated and then the average divergence vector is formed. This vector is considered as a classification sign for the class. The experiments were performed both in gray scale and HSV palette (component wise) and showed that for different classes average divergence vectors are substantially different in each color component. The method may be successfully applied for classification of biomedical preparation images.
Search Articles
Keywords: Direct multifractal transform, discrete probability distribution, Kullback-Leibler divergence, texture analysis
Pages: 194-202
WSEAS Transactions on Biology and Biomedicine, ISSN / E-ISSN: 1109-9518 / 2224-2902, Volume 11, 2014, Art. #25