TEXTURE CLASSIFICATION BY CENTER-SYMMETRICAL AUTOCORRELATION, USING KULLBACK DISCRIMINATION OF DISTRIBUTIONS

We propose a new method of texture analysis and classification based o n a local center-symmetric covariance analysis, using Kullback (log-li kelihood) discrimination of sample and prototype distributions. Featur es of our analysis are generalized, invariant, local measures of textu re having center-symmetric patterns, which is characteristic of many n atural and artificial textures. We introduce two local center-symmetri c auto-correlations, with linear and rank-order versions (SAG and SRAC ), together with a related covariance measure (SCOV) and variance rati o (SVR). All of these are rotation-invariant, and three are locally gr ey-scale invariant, robust measures. In classification experiments, we compare their discriminant information to that of Laws' well-known co nvolutions, which have specific center-symmetric masks. We find that o ur new covariance measures, which can be regarded as generalizations o f Laws' measures, perform better than Laws' approach despite their mea sure of texture pattern and grey-seal.