Take the 1-D DFT of one image, first row-wise 1-D DFT and followed by column-wise 1-D DFT to create the resulting F1(u,v) .
No products are associated with this question.
It actually doesn't matter, but you can do this
X = randn(8,8); Y = fft2(X); % ok let's first take the DFT along the rows rowdft = fft(X,,2); % now take it along the columns coldft = fft(rowdft,,1); max(max(abs(coldft-Y)))
See the maximum difference in absolute value is only on the order of 10^(-15)