denoise {EBImage}R Documentation

Noise removal, blurring and smoothing or images

Description

This set of functions allows for the removal of noise, blurring and smoothing of images. The functions operate of images in any image mode. The functions and the corresponding help descriptions are ported from ImageMagick, see the reference below.

Usage


  # Noise removal:
  ## S4 method for signature 'Image':
  denoise(x, r=0, ...)
  ## S4 method for signature 'Image':
  mediansmooth(x, r=2, ...)
  ## S4 method for signature 'Image':
  despeckle(x, ...)

  # Sharpening images:
  ## S4 method for signature 'Image':
  sharpen(x, r=0, s=0.5, ...)
  ## S4 method for signature 'Image':
  umask(x, r=0, s=0.5, amount=5, t=2, ...)

  # Blurring images:
  ## S4 method for signature 'Image':
  blur(x, r=0, s=0.5, ...)
  ## S4 method for signature 'Image':
  gblur(x, r=0, s=0.5, ...)

  # Adding noise to images:
  ## S4 method for signature 'Image':
  noise(x, type="G", ...)

Arguments

x An object of Image.
r A numeric value for the radius of the pixel neighbourhood. Passing 0 enables automatic radius selection, default.
s A numeric value for the standard deviation of the Laplacian (sharpen) or Gaussian (umask, blur, gblur), in pixels. For reasonable results, in most functions r must be larger than s.
amount A numeric value for the percentage difference between the original and the blurred image that is added back into the original in the un-sharp mask algorithm.
t A numeric value for the threshold in pixels needed to apply the amount in the un-sharp mask algorithm.
type The type of noise to add. Supported noise types are: Uniform, Gaussian (default), Multi, Impulse, Laplace and Poisson. The value can be specified by one letter. Case insensitive.
... Reserved.

Details

despeckle reduces the speckle-type, single-pixel, noise.

mediansmooth smooths the noisy image by replacing each pixel by a median of pixel values in taken over the neighbouring as defined by radius.

blur, gblur produce a blurred image. The blur method differs from the Gaussian blur, gblur, in that it uses a separable kernel which is faster but mathematically equivalent to the non-separable kernel.

sharpen, umask sharpen an image. umask uses the un-sharp mask algorithm, in which the image is convolved with a Gaussian operator of the given radius and standard deviation, s.

Value

A transformed image in an object of Image.

Author(s)

Oleg Sklyar, osklyar@ebi.ac.uk, 2005-2007

References

ImageMagick: http://www.imagemagick.org.

See Also

Image, normalize

Examples

  w <- 120
  a <- Image((0:(w^2))/w^2, c(w,w))
  if ( interactive() ) display(a)
  b <- normalize(noise(a) * 0.1)
  if ( interactive() ) display(b)
  dn <- despeckle(b)
  if ( interactive() ) display(dn)
  bl <- blur(dn, 4, 2)
  if ( interactive() ) display(bl)

[Package EBImage version 2.6.0 Index]