摘要
文章提出了一种MR图像增强的新算法。首先由多阈值分割算法将图像分割为不同的区域,然后统计出各区域的灰度均值和方差,由这两个特征构造各区域的累积指数非线性变换,并分别对相应区域进行增强。实验表明,用该算法增强后的MR图像在视觉上增强了图像对比度,加强目标的细节,还可以去除部分噪声,在实现上缩短了CPU时间。
In this paper,a new algorithm for MR image enhancement is proposed.First the image segmentation based on muhilevcl is carried out.Then the mean and variance of each segmented regions is calculated and used for constructing the accumulating index transformations,which is used to enhance the corresponding region.Compared to CLAHE,this algorithm effectively achieves the balance between the edge enhancement and noise retraining while it enhances the MR image luminance contrast with less CPU time,
出处
《计算机工程与应用》
CSCD
北大核心
2006年第16期29-31,52,共4页
Computer Engineering and Applications
基金
国家自然科学基金资助项目(编号:60302012)
宁波市科技攻关计划资助项目(编号:2005B100016)
关键词
图像增强
多阈值分割
非线性变换
Otsu阈值算法
累积指数变换
Image enhancement,multilevel segmentation,nonlinear transformation,Otsu threshold algorithm,accumulating index transformation
