Classifying image pixels into shaped, smooth, and textured points

摘要

An image is generally perceived by its contents of shapes, colors, and textures. In this paper, a novel algorithm based on fuzzy reasoning methods is proposed to classify each pixel of a gray-valued image into a shaped, a textured, or a smooth feature point. The method can effectively integrate the multi-scaled observations of an image, and the perceptually three-class classifying model can generally be used as a front-end for characterizing the content of an image.