Color image segmentation and understanding through connected components

A new definition of connected components for color images, called (r,δ) connected components, is proposed. This definition has developed the connected components theory. The new definition is applied to the segmentation and understanding of color images. Experiment shows that the segmentation through (r,δ)-connected components can give meaningful, reasonable and effective partition of color images. The (r,δ)-connected components analysis of color images is shown to be a good tool for the segmentation and understanding of color images