I am trying to remove low frequency bands from an image after transforming it using 2D fast fourier. I can identify individual components of the transformed image, but cannot yet work out how to filter only components higher than a certain frequency threshold and then reconstruct the original image using only those frequencies.
Say, for example, I have the following 5 x 5 image matrix, x:
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
I subtract out the mean and compute the FFT: x = x - mean(mean(x)); X = fft(x);
But then how would I filter only frequencies of X above a a threshold, t, and reconstruct the image using only those frequencies to give Y?