IMPROVED WHALE OPTIMIZING ALGORITHM BASED ON PSEUDO OPPOSITION-LEARNING AND DIFFERENTIAL EVOLUTION AND APPLICATION OF IMAGE SEGMENTATION

Aimed at the shortcomings of high computation cost and poor accuracy of the multi-level thresholding image segmentation, an improved whale optimization maximal entropy threshold image segmentation algorithm based on differential evolution and pseudo opposition-learning is proposed. In order to improve the optimization precision and the convergence rate of traditional WOA, we introduced the pseudo opposition-learning and chaos Tent map to generate the initial population, which could promote the population diversity and the quality of initial solutions, and expand the leading effect of elite individuals. The differential evolution was applied to enhance the global search of the population, which could overcome the shortcoming of falling into a local optimum at later iterations. Based on these works we achieved an improved algorithm OLDWOA. Using the maximal entropy as fitness function, OLDWOA was applied to search the optimal multi-level threshold group of image segmentation, in order to determine the optimal segmentation thresholds. Using classic images to construct image segmentation experiments, we compared some index such as the computational efficiency, the peak signal-to-noise ratio (PSNR), the structural similarity (SSIM) and the feature similarity (FSIM). The obtained experimental results verify that our algorithm can obtain a higher segmentation accuracy and a higher segmentation efficiency than the same kind.