Perform a 2D Fourier Transform in Matlab without using the inbuilt function

Hello Community,

Perhaps this is a silly one.

I'm trying to get the Fourier transform of an image using matlab, without relying on the fft2() function. As you'll see, I've tried taking the transform in three ways to compare the result but I'm unable to match the result with that obtained from the inbuilt function. Could someone please help me out here. I'm sure there is something simple that I am missing. Thank you in advance.

The code I'm using is as follows -

if true
  % code
a=imread('sq1.jpg');
Mat = double(a);  
[rows,columns]=size(Mat); 
A = zeros(rows,columns);
B = zeros(rows,columns);
for r=1:rows 
  for c=1:columns
      A(r,c)= A(r,c)+(Mat(r,c)*exp(((-1j)*2*pi*(r-1)*(c-1))/rows));
  end
end
for c=1:columns 
  for r=1:rows
      B(r,c)=(B(r,c)+(A(r,c)*exp((-1i)*2*pi*(r-1)*(c-1)/columns)));
  end
end
figure(1), imshow(B)
% Using the inbuilt function
a=imread('sq1.jpg');
Mat = double(a);
[rows,columns]=size(Mat);
for r=1:rows
  ffta(r,:)=fft(Mat(r,:));
end
for c=1:columns
  fftb(:,c)=fft(ffta(:,c));
end
figure(2), imshow(fftb)        % appliying 1d fft sequentially

 fftd=fft2(a);
 figure(3), imshow(fftd)      %applying 2d fft function
  end
end

[Alex Bjurstrom]

Did you ever figure this out? Wondering the exact same thing.

Answer 1

[Mustafa Rifat]

Could you solve this? I am looking for the same thing!