plot_feature.m

amer s noise level estimation implementation on MATLAB

clear all;
close all;
clc;


id = 60;
std = 15;
w_size = 5;
%==============================
% image read
%==============================

%noisy image x+w
s1 = 'E:\KODAK_PNG\noise\';
s2 = sprintf('img%02d', id);
s3 = '\';
s4 = sprintf('noise_std%02d',std);
s5 = '.png';    

originalfile = strcat(s1,s2,s3, s4,s5); 
a = imread(originalfile,'png');
noise = (double(uint8(a)));
noise_img = noise; 

s4 = sprintf('noise_std%02d',0);
originalfile = strcat(s1,s2,s3, s4,s5); 
a = imread(originalfile,'png');
ori = (double(uint8(a)));
ori_img = ori; 
%===========================================================================
% Masking operation for finding homogeneous blocks
%---------------------------------------------------------------------------
      
Cnt = 0;
img_width = size(noise,1);
img_height = size(noise,2);
center = median(1:w_size);

%     r1 = 0.3;%0.98;
%     r2 = 0.6;%0.01;
%     r3 = 0.1;%0.01; 
    
    r1 = 0.2;%0.98;
    r2 = 0.5;%0.01;
    r3 = 0.3;%0.01;    
    
    
    ratio = 0.0001;
    y_term = 125;    
    
    I_noise = uint8(r1 *noise_img(:,:,1) +  r2*noise_img(:,:,2) +  r3*noise_img(:,:,3));
    I_ori = uint8(r1 *ori_img(:,:,1) +  r2*ori_img(:,:,2) +  r3*ori_img(:,:,3));
        
    mI_noise = ratio.*double(I_noise) + y_term;
    mI_ori = ratio.*double(I_ori) + y_term;
    
    alpha = mI_noise - double(I_noise);
    alpha0 = mI_ori - double(I_ori);

    m_noise_R = noise_img(:,:,1) + alpha;
    m_noise_G = noise_img(:,:,2) + alpha;
    m_noise_B = noise_img(:,:,3) + alpha;
    m_noise = noise;
    
    m_ori_R = ori_img(:,:,1) + alpha0;
    m_ori_G = ori_img(:,:,2) + alpha0;
    m_ori_B = ori_img(:,:,3) + alpha0;
    m_ori = noise;    
       
    m_noise(:,:,1) = m_noise_R;
    m_noise(:,:,2) = m_noise_G;
    m_noise(:,:,3) = m_noise_B;

    m_ori(:,:,1) = m_ori_R;
    m_ori(:,:,2) = m_ori_G;
    m_ori(:,:,3) = m_ori_B; 
        
    m_noise_img = double(m_noise);
    m_ori_img = double(m_ori);
    
    %ori = m_ori;
    %noise = m_noise;
    
for i=1:w_size:img_width-w_size %x
    for j=1:w_size:img_height-w_size %y

        Cnt = Cnt+1;
        x_loc = i:i+w_size-1; 
        y_loc = j:j+w_size-1;
        
        rw(Cnt) = std2(noise(x_loc,y_loc,1));
        or(Cnt)  = std2(ori(x_loc,y_loc,1));
        wr(Cnt)  = sqrt(rw(Cnt)^2-or(Cnt)^2);
        
        mrw(Cnt) = std2(m_noise(x_loc,y_loc,1));
        mor(Cnt)  = std2(m_ori(x_loc,y_loc,1));
        mwr(Cnt)  = sqrt(mrw(Cnt)^2-mor(Cnt)^2);
        
    end
end
figure;plot(or,wr,'x');grid
xlabel('standard deviation of original image');
ylabel('standard deviation of AWGN');
title(['image id='  ,num2str(id),', window size='  ,num2str(w_size),  'given std =', num2str(std)]);

figure;plot(mor,mwr,'x');grid
xlabel('standard deviation of original image');
ylabel('standard deviation of AWGN');
title(['image id='  ,num2str(id),', window size='  ,num2str(w_size),  'given std =', num2str(std)]);