量子蚁群模糊聚类算法在图像分割中的应用

针对模糊C-均值算法对初始值的依赖,容易陷入局部最优值的缺点,本文提出将量子蚁群算法与FCM聚类算法结合,首先利用量子蚁群算法的全局性和鲁棒性以及快速收敛的优点确定图像的初始聚类中心和聚类个数,再将所得结果作为FCM聚类算法的初始参数,然后用FCM聚类算法对医学图像进行分割。实验结果表明,该方法有效解决了FCM算法对初始参数的依赖,克服了FCM算法及蚁群算法容易陷入局部极值的的缺点,而且在分割速度和精度上得到了较大提高。

Image Segmentation Based on Quantum Ant Colony Fuzzy Clustering Algorithm

Fuzzy C-Means algorithm is dependent on the initial value, resulting in easy to fall into the disadvantage of the local optimum value. A combination of quantum ant colony algorithm and FCM clustering algorithm is put forward. Firstly, the original center and numbers of cluster of the image are determined by using global type, robustness and advantages of fast convergence of quantum ant colony algorithm. Secondly, the obtained results are taken as the initial parameters for FCM clustering algorithm, and then the medical image is divided by using FCM clustering algorithm. It is proved that the method has reduced the dependence of FCM clustering algorithm on initial parameters effectively， overcome the shortcomings of easy falling into the local minimum of both algorithms，and greatly improved dividing speed and accuracy, which is simulated by real experiment.