How can I blur an image pixel by pixel, given an input image of 512 by 512 pixel

A patch of the image size 16 by 16 is blurred first leaving the others unchanged, then another step of blurring 32 by 32 leaving the others unchanged. I have started using the Gaussian filter but I don't know how to blur by each pixel on the same image

im=imread('Sharon2.jpg');
y=rgb2gray(im);
z=imresize(y,[512 512]);
x=imrotate(z,-270);
figure,
imshow(x);

Answer 1

[Image Analyst]

Not exactly sure what you want to do but look into using fspecial() and either conv2() or imfilter().

[Oluwamayowa Sanni]

The point is I have to use it for different pixel at a time not the whole image at once.

[Image Analyst]

Try this:

% Define blur kernel
hsize = 17;
sigma = 5; % Whatever
kernel = fspecial('gaussian', hsize, sigma)
% Read in sample image.
grayImage = imread('cameraman.tif'); % Sample image.
% Get a window around the target pixel that we
% will consider when blurring.
row1 = 100 - floor(hsize/2);
row2 = row1 + hsize - 1;
col1 = 150 - floor(hsize/2);
col2 = col1 + hsize - 1;
subImage = double(grayImage(row1:row2, col1:col2));
% Blur the pixel at row 100, column 150.
tempImage = subImage .* kernel
newPixelValue = uint8(sum(tempImage(:)));
% Put blurred value back into original image
grayImage(100, 150) = newPixelValue;

[Oluwamayowa Sanni]

So I do I take the blurred image through the image itself since I have to do it pixel by pixel

[Image Analyst]

No, because you said you don't want to blur the entire image. You said you wanted to blur just one pixel. If you scanned that kernel across the entire image, blurring each pixel one at a time pixel-by-pixel, you'd be blurring the entire image, which you said you definitely did not want to do. That would essentially be the same as using imfilter() and conv2() like I originally suggested but which you already rejected.

[Oluwamayowa Sanni]

let me just write you the question as it was asked...First, Resize the image to size of 512*512. Next, choose a patch of the image of size 16*16. After that, blurred the image patch. Then add a new portion to the patch and make it 32*32 where as the previous patch is being blurred. Continue this process till the last patch of the image.

[Image Analyst]

So, this is homework. You forgot to mention that or tag it, so I tagged it as homework above. So since it's homework, I can't just hand over sliding filter code for you since I have a feeling your instructor wanted you to figure it out yourself rather than just get the code from me. So just put the code I did give you into a loop over all rows and all columns.

for col = 1 : columns
    for row = 1 : rows
        % code to filter one pixel at location (row, col)
    end
end

[Oluwamayowa Sanni]

thank you but it is quite a challenge for me because this is my first time with learning codes and my deadline for submission is this night..I have tried all you told me but still not getting it done..Please

[Image Analyst]

Did you try putting my first code I gave you into the second code like I suggested?

% Define blur kernel
hsize = 17;
sigma = 5; % Whatever
kernel = fspecial('gaussian', hsize, sigma)
% Read in sample image.
grayImage = imread('cameraman.tif'); % Sample image.
[rows, columns, numberOfColorChannels] = size(grayImage)
w = floor(hsize/2);
for col = w+1 : columns - w
  for row = w+1 : rows - w
    % code to filter one pixel at location (row, col)
    % Get a window around the target pixel that we
    % will consider when blurring.
    row1 = row - w;
    row2 = row1 + hsize - 1;
    col1 = col - w;
    col2 = col1 + hsize - 1;
    subImage = double(grayImage(row1:row2, col1:col2));
    % Blur the pixel at row, column.
    tempImage = subImage .* kernel;
    newPixelValue = uint8(sum(tempImage(:)));
    % Put blurred value back into original image
    grayImage(row, col) = newPixelValue;   
  end
end
imshow(grayImage, []);
title('Blurred Image', 'FontSize', 30);
% Enlarge figure to full screen.
set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
% Give a name to the title bar.
set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')

Now, what are you going to do to make this assignment your own?

[Oluwamayowa Sanni]

Yes I did..

grayImage = imread('Sharon2.jpg'); % Sample image.
z=imresize(grayImage,[512 512]);
x=imrotate(z,-270);
y=rgb2gray(x);
figure,
imshow(y);
% Define blur kernel
hsize = 17;
sigma = 5; % Whatever
kernel = fspecial('gaussian', hsize, sigma);
% Read in sample image.
grayImage = imread('Sharon2.jpg'); % Sample image.
z=imresize(grayImage,[512 512]);
x=imrotate(z,-270);
y=rgb2gray(x);
figure,
imshow(y);
% Get a window around the target pixel that we
% will consider when blurring.
row1 = 100 - floor(hsize/2);
row2 = row1 + hsize - 1;
col1 = 150 - floor(hsize/2);
col2 = col1 + hsize - 1;
subImage = double(grayImage(row1:row2, col1:col2));
% Blur the pixel at row 100, column 150.
tempImage = subImage .* kernel;
newPixelValue = uint8(sum(tempImage(:)));
% Put blurred value back into original image
grayImage(100, 150) = newPixelValue;
figure,
imshow(tempImage);
% for col = 1 : columns
%     for row = 1 : rows
%         % code to filter one pixel at location (row, col)
%     end
% end