Abstract
The paper proposes new algorithms for combining superpixel regions into segments. The main idea is that when combining super pixels, firstly, a strategy is used in which the segment is grown from neighboring areas as long as the conditions for combining are met, and secondly, when combining areas, the applied information quality measure should not increase. Three algorithms based on this strategy are proposed, which differ in the conditions for making a decision on combining superpixels. A computational experiment was carried out on test images. The experiment showed that the proposed algorithms make it possible to speed up the segmentation process compared to the procedure used, with acceptable losses of information quality measures of the obtained partitions.