Reconstruction of Bilateral Filtering Based on NLM
Hui Wang, Liguo Shuai, Wenzhe Zhang, Huiling Chen · 2022
In order to improve the noise reduction effect of bilateral filtering, a new bilateral filtering algorithm incorporating the NLM algorithm is proposed. Considering that NLM focuses on the similarity between the location of noise reduction point and the location of adjacent points, NLM is used instead of gauss filtering to carry out a preliminary noise reduction, so that the filtering framework for noise reduction is determined. Then, based on the magnitude of the weighting coefficients of NLM and pixel greyscale obtained in the previous step, a matrix of weighting coefficients is obtained and analyzed to determine the size of the parameters in order to obtain better noise reduction. The experimental results show that the PSNR value is 2 ~ 3dB higher than that of the bilateral filter or NLM algorithm in the image with large gray difference, which shows that the reconstructed bilateral filter has obvious improvement compared with the traditional bilateral filter in the image with large gray difference.