Algorithms for Generating Image Segmentation Maps from Superpixels Based on a Combined Information Quality Measure

New algorithms for merging superpixels into segments using a greedy strategy and a combined quality measure including two components such as the information redundancy and variation of information were proposed. The greedy strategy was previously applied by the author to speed up image segmentation under the information redundancy minimum condition, and the two-component information quality measure was used in the problem of combining different segmentation maps. The simultaneous application of the greedy strategy with the combined quality measure in the new algorithms is aimed at accelerating the generation of segmentation maps and improving their quality through a compromise between the requirements of minimizing the number of informationally important segments and minimizing the information discrepancy between original images and generated partitions. A computational experiment on test images has demonstrated that the proposed algorithms can speed up the segmentation process as compared to the information redundancy minimization based method applied by the author earlier and improve the information characteristics of generated segmentation maps.