中文摘要：

在整个图像块像素灰度值向量空间中，非局部均值（nonlocal means，NLM）算法度量像素间的相似性不仅计算复杂度高，而且当噪声存在时还不能准确地计算出像素间的相似性权重值，影响了对图像冗余性质的利用，使得去噪结果图像对比度和清晰度低．针对NLM算法的这一缺陷，利用离散余弦变换（discrete cosine transform，DCT）的低数据相关性和高能量紧致性，将DCT与NLM算法相结合，对图像块进行DCT，并在DCT低频系数子空间内度量像素间的相似性．实验结果表明，与NLM算法相比，该方法能够在保护图像结构信息、对比度和清晰度的前提下更有效地去除噪声，峰值信噪比值一般可以提高1dB以上，运行时间不到NLM算法的1／10．

英文摘要：

Nonlocal means （NLM） has been becoming one of the most useful tools for image denoising. However, the computational cost is high due to the fact that calculation for similarity weights is performed in a full space of neighborhood patches. In addition, the calculation for its similarity weights has limited accuracy against noise when the noise standard deviation is large. In order to handle abovementioned problems, we introduce a novel image denoising algorithm that integrates discrete cosine transform （DCT） into the NLM method, motivating from promising characteristics of DCT, such as low data correlation and high energy compaction, etc. Our novel NLM-DCTS algorithm is capable of improving quality and reducing the computational cost. Compared with NLM algorithm, the PSNR value can be improved more than I dB by NLM-DCTS, and the running time can be reduced to 1/10 of NLM＇s.