基于邻域信息的多层次图像模糊算法

Multi-level image fuzzy enhancement based on neighborhood

摘要:

为了提高图像识别的准确率,提出了一种基于模糊理论的多层次图像增强算法.首先利用图像的统计特性对图像进行初步模糊划分,然后采用非线性连续阶梯状隶属度函数对图像进行模糊化,最后通过利用邻域特征自适应调节增强系数,对各模糊子集进行非线性增强处理以实现局部特征增强.通过与经典方法进行实验比较,显示出该增强算法能够在保存同质区内主要纹理特征的同时增强区域之间的对比度,增强了图像层次感,尤其对于医学图像,能改善图像质量,有助于提高医生临床诊断的有效性.

Abstract:

A multi-level grayscale-image enhancement method based on fuzzy technique is proposed to improve the contrast between homogeneous areas while saving textures.Nonlinear continuous stepped function based on statistical properties of the image is applied as membership function in fuzzy processing.The nonlinear enhancement is implemented by combining enhancement on each fuzzy sets.Comparing with classical methods,the results show that the image has higher contrast,enhance fine detail of image feature without enhancement,and it is better to representing the medical image processing,