Neural Networks for Textured Image Processing

Citations of this article
Mendeley users who have this article in their library.
Get full text


We review key conventional and neural network techniques for processing of textured images, and highlight the relationships among different methodologies and schemes. Texture, which provides useful information for segmentation of scenes, classification of surface materials and computation of shape, is exploited by sophisticated biological vision systems for image analysis. A brief overview of biological visual processing provides the setting for this study of textured image processing. We explain the use of multiple Gabor filters for segmentation of textured images based on a locally quasimonochromatic image texture model. This approach is compared to the functioning of localized neuronal receptive fields. Cooperative neural processes for perceptual grouping and emergent segmentation are reviewed next, and related to relaxation labelling. The recently developed SAWTA neural network for texture-based segmentation is then presented. Finally, techniques for describing and processing texture as a constrained optimization problem are outlined. © 1991, Elsevier B.V.




Ghosh, J., & Bovik, A. C. (1991). Neural Networks for Textured Image Processing. Machine Intelligence and Pattern Recognition, 11(C), 133–154.

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free