IMAGE SEGMENTATION BASED ON THE ANALYSIS OF DISTANCES IN AN ATTRIBUTE SPACE
P. A. Chochia
Institute for Information Transmission Problems, Russian Academy of Sciences, Bol’shoi Karentyi Pereulok 19, Moscow, GSP-4, 127994 Russia
Keywords: image processing, recognition, segmentation
Abstract
The image segmentation problem (partitioning into uniform regions) that uses brightness, color, and texture differences is considered. The criterion of uniformity is the estimation of the proximity of points in the combined attribute space. A metric is proposed for this space. This goal is achieved with a hierarchical algorithm based on the analysis of distances in the attribute space.
|
