An improved non-local means image denoising algorithm

Abstract

Non-Local means (NL-Means) algorithm is an effective denoising algorithm, but very dependent on the measured non-local similarity image blocks. In order to obtain as many similar image blocks as possible, we propose the Discrete Fourier Transform of local Gabor feature (LGF-DFT) which is rotation invariant and noise robust to measure the image blocks similarity, and fully utilize structural information of image patches. In addition, we optimize the weighted kernel function and the filter coefficient. Experimental results prove that the performance of NL-Means algorithm can be significantly improved by using the proposed similarity measure.