基于动态自适应蚁群算法的MRI图像分割

MRI图像分割在医学图像分析中具有极其重要的理论和应用价值.蚁群算法是一种具有离散性、并行性、鲁棒性和模糊聚类能力的进化方法.对目标边界模糊、目标灰度不均匀及目标不连续等情况的图像(如医学图像)分割,蚁群算法是一个比较好的选择.本文针对基本蚁群算法容易出现早熟和停滞现象的特性,提出了一种动态自适应蚁群算法,通过自适应的初始聚类中心调整策略和动态更新局部信息素浓度,使其收敛性和稳定性有一定的提高.实验证明改进的蚁群算法能够有效地分割MRI图像.

Segmentation of MRI Based on Dynamic and Adaptive Ant Colony Algorithm

Segmentation of MRI is very important in medical image analysis.Ant colony algorithm(ACA)is a kind of discrete,parallel and robust evolutionary algorithm with fuzzy clustering ability.To segment targets with blurry edges,intensity non-uniformity and discontinuity(such as medical images),ACA approach is a good choice.A dynamic and adaptive ant colony algorithm is presented in accordance with the defect of early variety and stagnation.The contribution of the algorithm includes an adaptive stratety of primary clustering center and a local updating for pheromone dynamically.Using this method can segment image fast and accurately.Experimental results show that the improved ACA is an effective MRI segmentation