function taking in an input image and different kernel size - matlab

Question

I have this filtering function that takes an input image, performs convolution using a given kernel, and returns the resulting image. However, I can't seem to work it out how to make it takes different kernel sizes.For example instead of pre-defined 3x3 kernel as below in the code, it could instead take 5x5 or 7x7. and then the user could input the type of kernel/filter they want(Depending on the intended effect). I can't seem to put my head around it. i'm quite new to matlab.
function [newImg] = kernelFunc(imgB)
img=imread(imgB);
figure,imshow(img);
img2=zeros(size(img)+2);
newImg=zeros(size(img));
for rgb=1:3
for x=1:size(img,1)
for y=1:size(img,2)
img2(x+1,y+1,rgb)=img(x,y,rgb);
end
end
end
for rgb=1:3
for i= 1:size(img2,1)-2
for j=1:size(img2,2)-2
window=zeros(9,1);
inc=1;
for x=1:3
for y=1:3
window(inc)=img2(i+x-1,j+y-1,rgb);
inc=inc+1;
end
end
kernel=[1;2;1;2;4;2;1;2;1]/16;
med=window.*kernel;
disp(med);
med=sum(med);
med=floor(med);
newImg(i,j,rgb)=med;
end
end
end
newImg=uint8(newImg);
figure,imshow(newImg);
end

Answer 1

I have commented the code and marked the places to change <--. kernel in this example is a column vector of 3*3=9 elements. Besides changing the kernel itself, you may also need to change the amount of zero-padding around the image. E.g., for a 5x5 kernel, you would need two rows and two columns of padding instead of one. Then update the inner loop I have marked "grab each pixel" to pull an area the size of the kernel (e.g., for x=1:5 and for y=1:5 for a 5x5 kernel). The actual convolution is unchanged.
Reminder: the function is taking a uint8 (0..255-valued) RGB image. window and kernel are double, so the trunc cuts any fractional part off before putting the new pixel value in uint8 newImg.
function [newImg] = kernelFunc(imgB)
img=imread(imgB);
figure,imshow(img);
img2=zeros(size(img)+2); % one extra column on each side, and one extra
% row top and bottom. <-- May need more padding.
newImg=zeros(size(img)); % the destination
% Pad the image with zeros at the left and top (transform img->img2)
for rgb=1:3
for x=1:size(img,1) %for each row
for y=1:size(img,2) %for each column
img2(x+1,y+1,rgb)=img(x,y,rgb); % <-- adjust per kernel size
end
end
end
% Process the padded image (img2->newImg)
for rgb=1:3
for i= 1:size(img2,1)-2 % for each row
for j=1:size(img2,2)-2 % for each column
% Build a row vector of the pixels to be convolved.
window=zeros(9,1); % <-- 9=kernel size
inc=1;
for x=1:3 % <-- grab each pixel
for y=1:3 % <-- under the kernel
window(inc)=img2(i+x-1,j+y-1,rgb);
inc=inc+1;
end
end
kernel=[1;2;1;2;4;2;1;2;1]/16; % <-- the kernel itself
med=window.*kernel; % start the convolution
disp(med);
med=sum(med); % finish the convolution
med=floor(med); % fit the pixels back into uint8.
newImg(i,j,rgb)=med; % store the result
end %for each column
end %for each row
end %for each color channel
newImg=uint8(newImg);
figure,imshow(newImg);
end