Theoretical-information quality model for image segmentation

A problem of image segmentation quality is considered. The problem is viewed as selecting the best segmentation from a set of images generated by segmentation algorithm at different parameter values. A segmentation quality model for selecting the best segmentation based on information redundancy measure is proposed. The developed technique was applied to SLIC and graph-cut segmentation algorithms. Computing experiment confirmed that the segmented image corresponding to minimum redundancy measure demonstrates suitable dissimilarity when compared with the original image. The segmented image which was selected using the proposed criterion, gives the highest similarity with the ground-truth segmentations.