Multi-level Thresholding Selection Based on Multi-Verse optimization With Levy Flight for Image Segmentation

Abstract

Multi-level thresholding(MLT) is a threshold-based image segmentation method. MLT is perhaps one of the most mainstream method of image segmentation because of its effective. For purpose of enhancing applicability and practicability of the MLT, a Levy Flight based multi-vese optimization(LFMVO) algorithm was proposed in this paper. Levy fight(LF) strategy is adopted to boost the global exploration ability of the multidimensional optimization problem and help the original algorithm to rapidly converge. It also can increase the vitality of the population of solution to let the algorithm avoid immature results. So as to evaluate the segmentation threshold of gray image, Kapurs entropy is selected as the segmentation basis. The performance and robustness of the proposed LFMVO algorithm were tested by using some randomly selected pictures from Berkeley image segmentation database (BSDS-500). LFMVO is compared with three other advanced meta-heuristic algorithm on the multilevel thresholding problem. The simulation results show the superiority of LFMVO and the final results give the proof that the LFMVO can enhance the image quality after segmentation, and further it be more stable than other algorithms.