A 3-D Texture Model Of Natural Scene Image

Abstract

A new 3-D texture model is developed by considering scene image as the superposition of a random texture image and a smooth shaded image. This smoothed surface is obtained by taking average of pixel intensities from natural image, and texture image can be generated by substracting the smoothed image from the original one. Then, Shape from-shading technique is applied to estimate the orientation of surface. For describing a random texture image, the long-correlation periodic model is chosen because of its good performance in texture synthesis and its ability to represent the coarseness and the pattern of the surface at the same time. Then, a projection technique is developed to deal with this particular random model, which has a non-isotropically distributed texture pattern. For estimating the parameters, a hybrid method which uses both the least square and the maximum likelihood estimates is applied directly to the given intensity function.