Classes
contrib/brl/bseg/sdet/sdet_denoise_mrf.h File Reference

a processor applying a binary mrf denoise algorithm More...

#include <vcl_vector.h>
#include <vnl/vnl_vector.h>
#include <sdet/sdet_denoise_mrf_params.h>
#include <vil/vil_image_resource.h>
#include <vil/vil_image_view.h>
#include <vnl/vnl_sparse_matrix.h>

Go to the source code of this file.

Classes

class  sdet_denoise_mrf

Detailed Description

a processor applying a binary mrf denoise algorithm

This algorithm selectively smooths the image based on a variance value at each pixel. The smoothing is carried out by a MRF with binary cliques, where the clique energy weights are controlled in part by the variance image.

Author:
J.L. Mundy - March 22, 2011

The mrf energy minimization is carried out with a closed form algorithm i.e. non-iterative as compared with belief propagation, which requires iterations of message passing. The algorithm is based in part on the paper: "Continuous MRF-based image denoising with a closed form solution," Ming Liu, Shifeng Chen, and Jianzhuang Liu Proc. IEEE 17th International Conference on Image Processing, 2010

    Modifications
     <none>
   

Definition in file sdet_denoise_mrf.h.