OWF: Non-local means and optimal weights for noise removal


The OWF software written in Matlab is a denoising algorithm to deal with the additive white Gaussian noise model. In the line of work of the Non-Local means approach, OWF is an adaptive estimator based on the weighted average of observations taken in a neighborhood with weights depending on the similarity of local patches. The idea is to compute adaptive weights that best minimize an upper bound of the pointwise L2 risk. In the framework of adaptive estimation, the “oracle” weights are optimal if triangular kernels are considered instead of the commonly-used Gaussian kernel. Furthermore, a way to automatically choose the spatially varying smoothing parameter is proposed for adaptive denoising. The implementation of the proposed algorithm (Matlab) is also straightforward and the simulations show that our algorithm improves significantly the classical NL-means and is competitive when compared to the more sophisticated NL-means filters both in terms of PSNR values and visual quality.

Software distribution

Matlab source code is freely ditributed. owe-software_mex.zip


Qiyu Jin, Ion Grama, Charles Kervrann, and Quansheng Liu. Non-local means and optimal weights for noise removal (2016).

Comments are closed.