Bayesian Network Based Texture Analysis and Classification

A novel Bayesian network（BN） texture model is proposed.Based on the model,a texture classification method is given.Pixel values in an image window are taken as a realization of a BN,and the structure and parameters of BN for each texture class are then studied.The conditional distribution probability of the pixels is used as the feature for texture classification.Experimental results show that the proposed model is effective for texture analysis and classification.