matlab-基于Split-Bregman迭代算法的医学图像去噪matlab仿真-源码

%% clc; clear; close all; warning off; addpath(genpath(pwd)); %% initialize parameters lamda=0.10; gamma=1; %% Shepp-Logan phantom N = 1024; % no. of pixels im =phantom(N); sigma= 0.10; %standard deviation of the noise % f = im+sigma*randn(size(im,1)); % add noise %% ------------------ figure;imshow(im); title('Original Image');colormap('jet'); % show the original image figure; imshow(f); title('Noisy Image'); colormap('jet'); % show the noisy image %% ------------------------------------- n=N^2; im= im(:); f=f(:); TOL =1e-3; % tolerance for split-Bregman iteration %% Split-Bregman TV denoising algorithm u_init=f; %initialize u(denoised image)with f(noisy image) [del,D1,D2,del_star,delta]=grad_operator(N); %calculate gradient matrix d= zeros(2*n,1); b= d; err= 1; it=1; %% Run the iteration tic; while err >TOL %% u subproblem [u_update,~]= cgs(speye(N*N)-delta, ... f+gamma*del_star*(d-b),1e-5,100); fprintf('iteration# %d\t',it); %% d subproblem d= d_subproblem_updated(del,u_update,lamda,gamma,b,n); %% update b b= b+del*u_update-d; % iter_term=norm((u_update-u_init),2) err=norm((u_update-u_init),2)./... %iteration tolerance norm(u_update,2); u_init=u_update; %update initial u fprintf('err=%d\n',err); it=it+1; end t.tv=toc; fprintf('\nReconstruction time (s)= %f\n',t.tv); fprintf('\nIteration stopped\n'); fprintf('\nRelative error is= %d\n', err); %% show denoised image u_update=reshape(u_update,N,N); figure; imshow(u_update);title('denoised image');