基于遗传模糊聚类的医学超声图像分割方法

医学超声成像技术以其实时性、无损性与廉价性等优点被广泛应用于医疗诊断，但由于其固有的斑点噪声和与组织相关的纹理特性使得医学超声图像的分割一直是一个难题。模糊C均值聚类算法（FCM）具有较强的抗噪声能力，能够较好地完成医学超声图像的分割任务，但其局限性在于对聚类中心的初值较敏感，当随机选取初始聚类中心时，很有可能使分割过程陷入局部极小，影响分割结果。利用遗传算法（GA）能够寻找全局最优解的特点，提出一种基于遗传算法寻找初始聚类中心的模糊聚类方法，应用于医学超声图像分割并取得了良好效果。

Medical ultrasound image segmentation method based on genetic algorithm and fuzzy clustering

Medical ultrasound imaging is widely used in medical diagnosis and treatment due to its characteristic of real-time,noninvasiveness and cheapness.However,the segmentation from medical ultrasound image is a difficult problem because of its intrinsic speckle noises and the tissue-related textures.Compared to other image segmentation methods,the Fuzzy C-means clustering algorithm(FCM) can fulfill the medical ultrasound image segmentation better,but it can easily be trapped in local optima due to its stochastic initialization of the clustering center.In this paper,a method of fuzzy clustering based on Genetic Algorithm(GA) is proposed and its robustness is more better because GA can obtain the global initialization of the clustering center.Experimental results show that the method can perform the segmentation from medical ultrasound images better.