ABSTRACT

The grey scale of an image is divided into 16 grades, s t s ale of the symbiotic matrix is 16×16. Then, calculate each feature vector of each sub-block by the above four formulas, when q is 0°, 45°, 90°, 135° . The symbiotic matrix of these four directions obtains e ergy, entropy, mom nt of inertia, and the correlation. Finally, the mea and the standard deviations of each parameter mCON, sCON, mASM, sASM, mRNT,sRNT, mCOR, sCOR, are regarded as the compo ents of t e texture feature vector. It is necessary to internally normalize the above eight components, since their physical meaning is different. The Gaussian normalization method is been used in this paper. A small amount of super or ul ra-mall value of elements has a little i pact on the distribution of the value of elements normalized.