Fuzzy Multilevel Image Thresholding Based on Improved Coyote Optimization Algorithm

Due to the computational complexity of multilevel image thresholding, Swarm Intelligence Optimization Algorithm (SIOA) has been widely applied improve calculation efficiency. Therefore, more and attention paid exploring application latest SIOA in segmentation. This article takes Otsu fuzzy entropy as objective functions, using Coyote (COA) for thresholds optimization selection, through median aggregation local neighborhood information then forms Fuzzy (FCOA), so that thresholding segmentation can be achieved end. To prevent COA algorithm from falling into optimum, this follows differential evolution strategy adopted by standard COA, number iterations construct scaling factor form Improved (ICOA). The experimental results show Kapur value aggregation-based ICOA(FICOA) achieves better quality. Compared with Grey Wolf Optimizer (GWO), Modified Quick Artificial Bee Colony Aggregation (FMQABCA) Discrete (FMDGWOA), FCOA FICOA have certain advantages visual effects PSNR, FSIM evaluation indices. Particularly compared GWO (also a wolf evolutionary algorithm), shows significant advantages.