Description

We focus on improving the denoising performance by the means of finding reliable candidate sets. We express patch-based locally optimal wiener (PLOW) filter as a summation of two related parts (NLM and residual filtering). The main problem in NLM denoising is to identify similar patches. The most time-consuming part of NLM is weight calculation, so a lot of methods are dominantly based on how to eliminate dissimilar patches before weighted averaging. The ease to increase the number of reliable candidates of noisy target patches was not concentrated. We propose clustering based classification , combined with PLOW for removing noise. Weiner filter provides the pre-processing for pre-classification. K-means clustering on LARK (Locally Adaptive Regression Kernels) features of the noisy image serves as the pre-classification for our filtering process. In the original NLM, all target patches have fixed candidate sets, which is either the whole image or the neighborhood centered at them. Finally calculate the average of weighted pixels.