滑动平均和改进权重函数的快速非局部平均图像去噪算法

非局部平均算法(NL-means)是一种有效的高斯噪声去除方法,由于其实现时效率低下,很难应用到实际中。针对非局部平均算法的低效率问题,提出一种快速的非局部平均去噪算法(FNLM)。首先,为了实现对算法的加速,采用滑动平均和权重对称技术。其次,算法在加速时一般会影响到去噪效果,为了使算法加速的同时保证去噪效果,提出一种改进的权重计算函数。最后,对新算法进行了一定的实验分析,实验结果显示提出的快速算法FNLM与原始的非局部平均算法相比,效率得到了很大提升,与其他的经典算法相比,在效率和效果上都非常有竞争力。

Fast non-localmeans for image de-noising on moving average and modified weight function

The non-localmeans(NL-means) algorithm provides a powerful framework for removing Gaussian noise. However,it is computationally impractical. In order to accelerate the algorithm,we use a moving average and weight symmetry in this paper. Speeding up the algorithm sometimes may reduce the quality,so we propose a modified weight function for calculating the weights. Finally,numerical results reveal that the proposed algorithm is faster than the original non-localmeans,and is also very competitive to most of the state-of-the-art algorithms in terms of both the PSNR and the subjective visual quality.