基于改进果蝇优化的模糊C聚类图像分割

Fuzzy C Clustering for Image Segmentation Based on Improved Fruit fly Optimization

针对传统的模糊C聚类算法的局限性,本文引入核函数来对目标函数做进一步优化,并在聚类算法的基础上引入果蝇优化和混沌的思想,提出一种新的图像分割算法。根据混沌运动的特性与果蝇优化算法进行合并,对模糊核聚类算法优化,进一步减小聚类算法对初始参数值敏感而陷入局部极值的负面影响。本文改进的算法能够达到良好的图像分割效果,同时具有较快的收敛性和良好的鲁棒性。

In view of the limitation of the traditional fuzzy C clustering algorithm , we introduc the kernel func- tion to polish up the ability to solve problems by objective function , such as the problem of noise sensitivity . this paper introduces the idea of optimization and chaos of the fruit fly based on the clustering algorithm , and then presents a new algorithm for image segmentation.This new algorithm use the character of Ergodicity and Random- ness by chaotic motion, atthe same time fruit fly optimization algorithm has a quick optimization speed and can quickly find the global optimal solution . Combine their features with each other can polish up the fuzzy kernel clustering algorithm to reduce the negative effect that the traditional algorithm was sensitive to initial value and fall into localoptimal value prematurely . Therefore , The new improved algorithm not only achieve a better effect of image segmentation , but also has fast convergence and robustness.