Discrimination of Natural Textures: A Neural Network Architecture.

摘要

A neural architecture for computing textural segmentation was synthesized from principles borrowed from neurophysiology. The underlying premise is that textural segmentation can be achieved by recognizing local differences in texture elements (texels). This approach differs from most of the previous work where differences in global, second-order statistics of the image points are used as the basis for segmentation. The architecture proposed in this paper consist of three major components: a feature extraction network; a local boundary detection network; and a higher-order texture discrimination network. Interactions between these networks result in the segmentation of the textured image.