摘要
为准确分割脑磁共振图像以帮助医护工作者进行诊断,提出一种改进的模糊C均值聚类与自适应非局部均值相结合的分割算法。算法对MR图像进行自适应NLM降噪处理,估计噪声图像方差,根据估计的方差为MR图像自适应地选择最佳搜素窗口大小;结合邻域像素的局部空间信息,在隶属度函数中引入条件空间变量;使用改进的条件空间FCM算法对降噪后的图像进行分割。实验结果表明,该算法在高噪声下仍能保持较高的分割精度,相比于传统算法具有更高的技术价值和应用前景。
In order to accurately segment brain magnetic resonance images to help medical workers diagnose,an improved segmentation algorithm combining fuzzy C-means clustering and adaptive non-local means is proposed.The algorithm performs adaptive NLM noise reduction on MR images,estimates the variance of noise images,and adaptively selects the optimal search window size for MR images according to the estimated variance.Combining the local spatial information of neighboring pixels,conditional spatial variables are introduced into membership function.An improved conditional space FCM algorithm is used to segment the denoised image.The experimental results show that the algorithm can still maintain high segmentation accuracy under high noise,and has higher technical value and application prospect than the traditional algorithm.
作者
陈凯欣
周安琪
蒋林华
CHEN Kaixin;ZHOU Anqi;JIANG Linhua(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;The 47th Institute of China Electronics Technology Group Corporation,Shenyang 110032,China)
出处
《微处理机》
2020年第1期27-32,共6页
Microprocessors
基金
国家自然科学基金资助项目(61775139)
上海市高校特聘教授(东方学者)岗位计划资助(15HJPY-MS02)
