About

Author

Thierry Guillemot <thierry.guillemot.work@gmail.com>, Julie Delon <julie.delon@parisdescartes.fr> and Agnès Desolneux <agnes.desolneux@cmla.ens-cachan.fr>

Copyright

(C) 2015 IPOL Image Processing On Line http://www.ipol.im/

Licence

GPL V3+

Copying and distribution of this file, with or without modification, are permitted in any medium without royalty provided the copyright notice and this notice are preserved. This file is offered as-is, without any warranty.

Overview

This source code provides an implementation of the “Parigi”: a patch-based approach to remove impulse-gaussian noise image, as described in IPOL.

The ‘bin/parigi’ program reads an PNG image, add impulse-gaussian noise and then apply the parigi denoising algorithm. The algorithm estimates the probability of impulse noise and extract similar patches. From them, it estimates for each pixels an histogram of all possible values. The denoised image is defined from the best mode of each histograms.

Only 8bit GRAY PNG images are handled.

Requirement

The code is written in UTF8 C++, and should compile on any system with an UTF8 C++ compiler.

The libpng and FFTW header and libraries are required on the system for compilation and execution. On Linux, just use your package manager to install it:

sudo apt-get install libpng
sudo apt-get install libfftw3-dev

For more information, see http://www.libpng.org/pub/png/libpng.html and http://www.fftw.org/.

Compilation

Simply use the provided makefile, with the command ‘make’. The makefile will produce a program called : ‘bin/parigi’.

It is possible to compile the program using OpenMP with the command ‘make OMP=1’.

The ‘parigi’ program is used to corrupted an original image with an impulse-gaussian noise and the apply the parigi restoration.

Usage

Parigi

Description

The ‘parigi’ program is used to corrupted an original image with an impulse-gaussian noise and the apply the parigi restoration. It takes 8 parameters with other optional:

bin/parigi input.png noisy.png denoised.png p_gaussian_noise p_impulse_noise two_step mixed_image verbose

input.png

input image.

noisy.png

resulting noisy image corrupted with an impulse-gaussian noise.

denoised.png

resulting parigi denoised image.

p_gaussian_noise

noise gaussian standard deviation [0. 255.0].

p_impulse_noise

probability of the impulse noise [0.0, 1.0].

two_step

apply parigi algorithm two times (allows to obtain best visual results) [0 or 1].

mixed_image

apply a criterium to mix the noisy and the denoised images (allow to obtain best visual results) [0 or 1].

verbose

activate the verbose mode

Usage examples

If you want to compute the parigi restoration of the file ‘input/barbara.png’ corrupted with a gaussian noise of standard deviation of 10 and an impulse noise with a probability of 40% by applying two steps with the mixed criterium without the verbose mode and saving the noisy image into ‘noisy.png’ and the denoised image into ‘denoised.png’, you can use the command:

bin/parigi input/barbara.png noisy.png denoised.png 10 0.4 1 1 0