A Graph Structure for Grey Value and Texture Segmentation

Abstract

An attempt for developing an unified method for grey value and texture segmentation was made. It makes use of a special graph structure (Feature Similarity Graph—FSG) which is based on a feature similarity criterion and a feature smoothing procedure applied in each layer of the network. Starting with grey value segmentation (the features are the pixel grey values) one obtains segments which, for textured images, represent texture elements (texels) or parts of texels and background, respectively. The texels can be described by certain features, namely position, orientation, size, grey value or color, and shape descriptors. Studying position and orientation, spatial frequency phenomena and important observations made by investigators of human perception, especially the Gestalt laws, can be explained. The highly parallel O(N) method can be applied also to the clustering of dot patterns.