Title :
Texture segmentation using Voronoi polygons
Author :
Tüceryan, Mihran ; Jain, Anil K.
Author_Institution :
Dept. of Comput. Sci., Michigan State Univ., East Lansing, MI, USA
fDate :
2/1/1990 12:00:00 AM
Abstract :
Textures are defined in terms of primitives called tokens. A texture segmentation algorithm based on the Voronoi tessellation is discussed. The algorithm first builds the Voronoi tessellation of the tokens that make up the textured image. It then computes a feature vector for each Voronoi polygon. These feature vectors are used in a probabilistic relaxation labeling on the tokens, to identify the interior and the border regions of the textures. Some experimental results are shown
Keywords :
computational geometry; pattern recognition; picture processing; Voronoi polygons; Voronoi tessellation; pattern recognition; picture processing; probabilistic relaxation labeling; texture segmentation; textured image; tokens; Computer science; Computer vision; Data mining; Focusing; Image segmentation; Labeling; Machine vision; Shape; Statistics; Surface texture;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on