An Effective Improvement on the Non-Local Means Filter for Removing the Mixture of Gaussian and Impulsive Noises
The non-local means (NL-means) denoismg filter preserves the fine structure and detail well in noisy images. However, the NL-means filter cannot remove impulsive noises effectively. In this paper, we propose a new denoising filter, which is an improvement of the NL-means filter. This new filter is based on the NL-means filter, which is modified with the median range of grey-values in the noisy image dynamically according to the impulsive noise density. An automatic tuning method is proposed to determine a range parameter and testing threshold. Experimental results show that our proposed method not only removes both impulsive and Gaussian noises effectively but also preserves edges and details of the structure well.
non-local means median filter impulsive noise Gaussian noise
Hong Liu Ning Pan Cihui Yang Chih-Cheng Hung
School of Computer Science and Technology, Huazhong University of Science and technology, Wuhan, Hub School of Computing and Software Engineering, Southern Polytechnic State University, 1 Marietta, Geo
国际会议
海口
英文
346-350
2011-07-15（万方平台首次上网日期，不代表论文的发表时间）