A Modified Moth-Flame Optimization Algorithm for Image Segmentation

As a renowned swarm-based approach, Moth-Flame Optimization (MFO) has been developed to address global optimization difficulties. It has been used to address a variety of advancement difficulties since its introduction. However, MFO may have difficulties in global optimal point and also suffer from slow convergence pace. In this paper, a modified MFO (m-MFO) is designed to address these issues by introducing a new parameter C and a weight W. The parameter C is utilized to boost exploration, whereas the weight W aids the algorithm in local search. The suggested approach is applied to the solution of 20 benchmark functions from both the unimodal and multimodal categories. Finally, the methodology is used to image segmentation using a multilayer thresholding approach. The comparison of the evaluated findings with alternative metaheuristic algorithms and convergence studies shows that the proposed m-MFO approach outperforms the basic MFO approach and several other methods.