摘要
数字X光影像仪(CR)图像对比度低、细节不清晰,要对其进行增强处理方能满足临床诊断的需要,而目前通用的CR图像增强算法对比度和噪声增强过度、丢失细节,为此提出了基于小波变换的CR图像增强法。该算法将CR图像进行小波变换分解,针对各子图像的特征,对高频细节部分采用非线性函数进行对比度增强,提高细节清晰度;对低频平滑部分采用反锐化掩模(UM)的方法对图像进行增强,以保证CR图像整体增强效果,最后进行小波变换得到增强后的CR图像。实验证明,该算法处理后的CR图像细节丰富,细节方差(DV)比通用增强算法提高了将近3倍,而且信噪比高,细节方差和背景方差(BV)之比(DV/BV)约为其他算法的3.5倍,具有良好的视觉效果。
To enhance OR image to improve its visual quality,an enhancement algorithm for CR image was proposed based on the multiresolution properties of wavelet analysis. The algorithm decomposes CR image into four sub-images,and uses different processing methods for the every sub-image enhancement. The sub-image with high frequency uses nonlinear function to enhance the detail definition, and then the unsharp masking（UM） method was used to enhance the sub-image with low frequency, at last gets the enhanced image by wavelet inverse transform. Experimental results demonstrate that the algorithm enhances OR image detail and the signal to noise ratio,the detail variance（DV） of the processed image is higher 3 times than the image enhanced by the general enhancement algorithms,and the value of DV/BV is higher 3.5 times than the other algorithms,while enhancing CR image with good visual effect.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2005年第8期989-992,共4页
Journal of Optoelectronics·Laser
基金
中国科学院二期创新资助项目(CO2E06Z)
