I want to remove noise using multi resolution SVD.

4 次查看（过去 30 天）

sufian ahmed 2017-6-6

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/343611-i-want-to-remove-noise-using-multi-resolution-svd

Here is my code . i want to remove noise using this multi resolution SVD. but it not works. After execution noise still exists. please someone modify this code so that it will remove noise Thanks.

Here is the algorithm which i want to implement.i got it from a research paper. i also include the paper here.

ference image (ground truth)

    imt = imread('eight.tif');
    imt= im2double(imt);
    imt=imresize(imt,[256,256]);
    figure; imshow(imt);
    J = imnoise(imt,'salt & pepper',0.1);
    figure, imshow(J);
  %apply MSVD
  [X1, U1] = MSVD(J);
  %apply IMSVD
  imf = IMSVD(X1,U1);
%%%%%%MSVD %%%%
function [Y,U] = MSVD(x)
% multiresolution SVD (MSVD)
% input-> x: image (spatial domain)
% outputs-> Y: one level MSVD decomposition of x
%           U: the unitary matrix (U in SVD)
% x=imresize(x,[256,256]);
[m,n] = size(x);
m = m/2; n = n/2;
A = zeros(4,m*n);
for j = 1:n
    for i = 1:m
        A(:,i + (j-1)*m) = reshape(x((i-1)*2+(1:2),(j-1)*2+(1:2)),4,1);
    end
end
[U,S] = svd(A);
U
S
T = U'*A;
Y.LL = reshape(T(1,:),m,n);
Y.LH = reshape(T(2,:),m,n);
Y.HL = reshape(T(3,:),m,n);
Y.HH = reshape(T(4,:),m,n);
%%%%IMSVD %%%%%
function[x1] = IMSVD(Y,U)
% inverse MSVD (IMSVD)
% Inputs-> Y: MSVD coefficients & U: unitary matrix (U in SVD)
% output-> x: image (spaitial domain)
[m,n] = size(Y.LL);
mn = m*n;
T1 = zeros(4,mn);
T1(1,:) = reshape((Y.LL),1,mn);
T1(2,:) = reshape((Y.LH),1,mn);
T1(3,:) = reshape((Y.HL),1,mn);
T1(4,:) = reshape((Y.HH),1,mn);
% figure,imshow(Y.LL,[]);
A1 = U * T1;  
x1 = zeros(m*2,n*2); 
for j = 1:n
    for i = 1:m
        x1((i-1)*2+(1:2), (j-1)*2+(1:2)) = reshape(A1(:,i+(j-1)*m),2,2);
    end
end
figure,imshow(x1,[]);