ABSTRACT
This entry reviews the basic concepts behind image segmentation (IS) algorithms. The approach that was followed consists of splitting any IS process into two distinct parts: definition of a homogeneity criterion and definition of an aggregation strategy. With regard to homogeneity, this entry mainly focuses on the segmentation of gray-level images by relying on the gray level of any single-image pixels. Aggregation methods are described by considering the four main paradigms usually adopted in the literature-postprocessing aggregation, split and merge, region growing, and Markov random fields.
