This paper introduces a novel denoising method called Non-Local Means (NLM). The method replaces the color of a pixel with an average of the colors of similar pixels, but these similar pixels do not need to be close to the target pixel. The authors present two implementations of the method: pixelwise and patchwise. The pixelwise implementation averages the colors of pixels within a fixed-size neighborhood, while the patchwise implementation averages the colors of patches within a larger neighborhood. Both methods use an exponential kernel to weight the contributions of similar patches. The paper also discusses the parameters and provides examples demonstrating the effectiveness of the NLM method in removing noise while preserving fine details. The source code and documentation are available online, with some files linked to a patent.This paper introduces a novel denoising method called Non-Local Means (NLM). The method replaces the color of a pixel with an average of the colors of similar pixels, but these similar pixels do not need to be close to the target pixel. The authors present two implementations of the method: pixelwise and patchwise. The pixelwise implementation averages the colors of pixels within a fixed-size neighborhood, while the patchwise implementation averages the colors of patches within a larger neighborhood. Both methods use an exponential kernel to weight the contributions of similar patches. The paper also discusses the parameters and provides examples demonstrating the effectiveness of the NLM method in removing noise while preserving fine details. The source code and documentation are available online, with some files linked to a patent.