作　　者： ; ; ;

机构地区： 江西理工大学信息工程学院

出　　处： 《计算机应用研究》 2012年第2期764-766,774,共4页

摘　　要： 为了确定医学图像的最佳灰度直方图熵,提出了一种基于改进演化算法的快速分割方法,能够自适应调整交叉和变异概率,既保证了种群的多样性,又克服了传统演化算法局部最优、收敛过快的缺点。搜索到的最佳阈值不仅比传统演化算法稳定性好,还有效地缩短了搜索时间,快速地实现了医学图像分割,而且分割后的图像可读性强。实验结果表明,该方法速度快、分割效果好。 In order to determine the best histogram entropy of medical image,this paper proposed a fast segmentation method based on improved evolutionary algorithm which could adjust the crossover and mutation possibility adaptively.It ensured the population diversity and overcame the local optimal and fast convergence of traditional evolutionary algorithm.The best thre-shold searched was not only more stability but also greatly reduced searching time,and rapidly implemented medical image segmentation,moreover,the image after segmentation had a strong readability.Experimental results show that the method is faster and has a better segmentation effect.

关 键 词： 医学图像分割 直方图 演化算法 熵 阈值

领　　域： [自动化与计算机技术] [自动化与计算机技术]