GSCAD (version 0.1.0)

gscad.denoise: Use GSCAD to denoise image

Description

Use the GSCAD method under dictionary learning framework to learn a proper sized dictionary and denoise image. The noisy image is split into m by m patches and a dictionary with each atom of size m is learned. The final denoised image is reconstruncted on the denoised patches.

Usage

gscad.denoise(I_noise, sigma, D0 = NULL, m, p0, c = 3.7, lambda = 0.05,
  maxrun = 20, maxrun_ADMM = 20, err_bnd = 1e-04, err_bnd2 = 1e-04,
  rho = 16, cor_bnd = 1, L = NULL)

Arguments

I_noise

The image to be denoised. In form of matrix.

sigma

Noise level.

D0

Initial dictionary. If D0 specified, m and p0 are not needed, otherwise D0 is evaluated as overcompleted DCT basis using function ODCT(m,p0). Either D0 or (m,p0) needs to be specified.

m

The size of the small patches to be split.

p0

Initial size of the dictionary.

c, lambda

Parameters for GSCAD.

maxrun

(optional) Maximun number of outer iterations to run. Default is 20.

maxrun_ADMM

(optional) Maximun number of iterations to run for updating dictionary using ADMM. Default is 20.

err_bnd

(optional) Stopping criterion for iterations. Default is 1e-4.

err_bnd2

(optional) Stopping criterion for updating dictionary

rho

(optional) Parameter for ADMM. Default is 16.

cor_bnd

(optional) When normalize dictionary, checking if the correlation of any two atoms are above the cor_bnd, one of the atom is removed. Default is 1.

L

(optional) This parameter controls the maximum number of non-zero elements in each column of sparsecolding A.Default is m.

Value

The learned dictionary dictionary, its size p and the denoised image fitted image.

Details

See https://arxiv.org/abs/1605.07870

Examples

# NOT RUN {
I = lena_crop #use a smaller image as an example
## add noise
sigma=20;
I_noise=AddNoise(I,sigma)
## denoising using GSCAD
# }
# NOT RUN {
out=gscad.denoise(I_noise,sigma,m=64,p0=100)
# }