改进型Q-NLM算法在医学图像去噪中的应用

Advanced Q-NLM algorithm application in medical image denoising

将像素离群度与传统非局部均值算法相结合提出一种改进型的非局部均值滤波Q-NLM算法,针对传统非局部均值算法不适应脉冲噪声去噪的问题,提出了像素离群度Q的概念,像素离群度Q值用于判断原像素与脉冲噪声的相似度,依据像素离群度Q值划分像素区域,调整不同区域阈值且引入中值滤波去除脉冲噪声点,进一步降低医学图片中混合噪声对滤波的影响。仿真实验结果显示,这种结合离群度与非局部均值算法在去除混合噪声的情况下,能较好提高图像信噪比,有效保留CT图像细节。

The pixel quotient is combined with traditional nonlocal means algorithm, putting forward a kind of improved nonlocal means filtering algorithm. That is Q-NLM. In allusion to the problem of the impulsive noise in the traditional nonlocal means algorithm, the concept of the pixel quotient Q is come up with. It is used for judging the similarity of the initial pixel and impulse noise, and according to the calculated Q to divide pixel areas, the median filter is brought in to re-move impulse noise points while adjusting different area threshold. Thus it can further reduce the influence of the mixed noise on filter in medical images. The relevant simulation experiment results show that the combination of the pixel quotient with traditional nonlocal means algorithm can improve the image PSNR, effectively keep the details of CT image while removing the mixed noise.