Maximum 2-D entropy image segmentation method based on improved sparrow algorithm

In order to improve the performance of the maximum 2-D segmentation, an image segmentation method based on improved sparrow algorithm (ITSSA) was proposed, which can decrease the amount of computation and shorten the time. Firstly, the reverse learning strategy and adaptive t-distribution variation were combined, while elite particles were introduced to expand the search range of the algorithm and to increase the local search ability of the algorithm in the later stage. Secondly, the firefly mechanism was used to perturb and mutate the optimal solution for the further increasement of the population diversity. Finally, the improved sparrow algorithm was used to find the maximum 2-D entropy of the image, and then the optimal threshold segmentation image was obtained. The results show that, the average running time of the proposed algorithm in the four images is 0.3695s, which is much lower than 1.7547s of the basic two-dimensional entropy algorithm and 5.7936s of the basic two-dimensional Otsu algorithm. The global search and local optimization ability of ITSSA, compared with the original sparrow algorithm, improves a lot, and the proposed segmentation method in this paper greatly shortens the traditional maximum 2-D entropy image segmentation method of running time. Apart from that, both the peak signal to noise ratio and the feature similarity index of this method increase, which has a certain application value.