Texture classification using noncasual hidden Markov models

The problem of using noncausal hidden Markov models (HMMs) for texture classification is addressed. In noncausal models, the state of each pixel may be dependent on its neighbors in all directions. New algorithms are given to learn the parameters of a noncausal HMM of a texture and to classify it into one of several learned categories. The efficacy of these algorithms for texture classification is determined by classification experiments involving both synthetically generated and natural textures. A comparison to recent results in autocorrelation modeling demonstrates that similar classification accuracy can be achieved using noncausal HMMs that learn fewer parameters