基于区域分割的自适应反锐化掩模算法

针对线性反锐化掩模法的缺点,提出一种基于区域分割的自适应反锐化掩模法,利用像素局部方差将输入图像划分为平滑区、中等对比度区和高对比度区,根据像素(x,y)所属的区域类型,自适应确定其局部灰阶变化率增益系数α(x,y)及期望的输出局部灰阶变化率H_d(x,y),并导出增强系数K(x,y)。对一幅胸片图像进行处理,并与其它几种反锐化掩模法的处理结果进行比较,结果显示该方法获得较好的边缘增强效果,且有较强的噪声抑制能力。

Adaptive unsharp masking method based on region segmentation

A adaptive unsharp masking method based on region segmentation is presented aiming at the defects of linear unsharp masking. The input image is divided into homogeneous areas, medium contrast areas and large contrast areas by the local variance of image pixels. Base on the area type to which image pixel (x,y) belongs, the local activity gain factor α(x,y) and the desired output local activity H_d(x,y) are determined adaptively, and the enhancement factor K(x,y) is derived. A X-ray chest image is processed with the proposed method and other unsharp masking methods. The results show that good capabilities of edge enhancing and noise suppressing are achieved by the proposed method.