image_denoise_nlmeans {image.DenoiseNLMeans} | R Documentation |
Image Denoising using Non-Local Means
image_denoise_nlmeans(image, sigma = 10, win = 1, bloc = 10, filter = 0.4)
image |
an object of class |
sigma |
sigma noise parameter (type numeric) |
win |
Half size of comparison window (type integer) |
bloc |
Half size of research window (type integer) |
filter |
filtering parameter (type numeric) |
an object of class magick-image
with the denoised image,
in the comments the algorithm arguments sigma, filter, window and bloc are provided.
Function is provided for scientific and education only. May be linked to the EP patent 1,749,278 by A. Buades, T. Coll and J.M. Morel.
For details on the algorithm, check the paper http://www.ipol.im/pub/art/2011/bcm_nlm.
library(magick) f <- system.file(package = "image.DenoiseNLMeans", "extdata", "img_garden.png") img <- image_read(f) img denoised <- image_denoise_nlmeans(img) denoised x <- image_noise(img, noisetype = "Poisson") x denoised <- image_denoise_nlmeans(x, sigma = 40) denoised denoised <- image_denoise_nlmeans(x, sigma = 10, win = 2, bloc = 10, filter = 0.4) denoised image_comment(denoised)