I needed to plot Histograms (or) Loss Information for original image and after restoration of blur image

%% Input Image
Ioriginal = imread('greyface.jpg');
imshow(Ioriginal)
title('Original Image')

% % Restore Motion Blur Without Noise
PSF = fspecial('motion',21,11);
Idouble = im2double(Ioriginal);
blurred = imfilter(Idouble,PSF,'conv','circular');
imshow(blurred)
title('Blurred Image')

%%Restoring blur image 
wnr1 = deconvwnr(blurred,PSF);
imshow(wnr1)
title('Restored Blurred Image')

%%Restore Motion Blur and Gaussian Noise
noise_mean = 0;
noise_var = 0.0001;
blurred_noisy = imnoise(blurred,'gaussian',noise_mean,noise_var);
imshow(blurred_noisy)
title('Blurred and Noisy Image')

wnr2 = deconvwnr(blurred_noisy,PSF);
imshow(wnr2)
title('Restoration of Blurred Noisy Image (NSR = 0)')

signal_var = var(Idouble(:));
NSR = noise_var / signal_var;
wnr3 = deconvwnr(blurred_noisy,PSF,NSR);
imshow(wnr3)
title('Restoration of Blurred Noisy Image (Estimated NSR)')

[Kartikeya]

I Have to get a plots or loss information based on original image vs Restoration of Blurred Noisy Image

Answer 1

[Image Analyst]

Why not use imabsdiff to get the difference between the original image and the restored image and then take the histogram of that difference image?

diffImage = imabsdiff(Ioriginal, wnr1);
histogram(diffImage);

[Kartikeya]

@Image Analyst I am getting error

Error using checkForSameSizeAndClass (line 8)

X and Y must have the same class.

Error in imabsdiff (line 42)

checkForSameSizeAndClass(X, Y, mfilename);

Error in Untitled3 (line 35)

diffImage = imabsdiff(Ioriginal, wnr1);

>> Untitled3

Error using checkForSameSizeAndClass (line 8)

X and Y must have the same class.

Error in imabsdiff (line 42)

checkForSameSizeAndClass(X, Y, mfilename);

Error in Untitled3 (line 35)

diffImage = imabsdiff(Ioriginal, wnr1);

[Image Analyst]

So make them both the same class, like uint8 or double.

[Kartikeya]

@Image Analyst Even though i am getting same error can you please change the code and send me.

[DGM]

Ioriginal = imread('cameraman.tif');

% Restore Motion Blur Without Noise
PSF = fspecial('motion',21,11);
Idouble = im2double(Ioriginal);
blurred = imfilter(Idouble,PSF,'conv','circular');

% Restoring blur image 
wnr1 = deconvwnr(blurred,PSF);

% assuming the abs difference is the thing you actually want
err = imabsdiff(Idouble,wnr1);
histogram(err)

% Restore Motion Blur and Gaussian Noise
noise_mean = 0;
noise_var = 0.0001;
blurred_noisy = imnoise(blurred,'gaussian',noise_mean,noise_var);

signal_var = var(Idouble(:));
NSR = noise_var / signal_var;
wnr3 = deconvwnr(blurred_noisy,PSF,NSR);

% assuming the abs difference is the thing you actually want
err = imabsdiff(Idouble,wnr3);
histogram(err)