Multiobjective Optimization Approaches in Image Segmentation – The Directions and Challenges

A new trend of problem formulation for image segmentation is touse multiobjective optimization approach in its decision makingprocess. Multiobjective formulations are realistic models for manycomplex optimization problems. In many real-life problems, objectives under consideration conflict with each other, andoptimizing a particular solution with respect to a single objective canresult in unacceptable results with respect to the other objectives. The aim of this paper is to provide a comprehensive review ofmultiobjective optimization in image segmentation problem. Themodels are classified according to relevant features, such as thevarious aspects of the optimization approaches used, various possibleproblem formulations, type of datasets modelled and the scope of theapplications. Through our analysis of the current state of theresearch, we diagnose some of the directions and challenges formodelling the image segmentation problem with multiple objectivescriteria. In this review, we consolidate the selected material in theliterature, including more than 80 studies dated 1990 or later