Frequency filtering is implemented by first accepting a list of parameters defining a frequency filter.When the filtering is applied, image is transformed to frequency domain using a 2 dimentional Fast FourierTransform (FFT) algorithm. In the transferred image, frequency in different range is filtered outaccording to the filter defined by user. Finally the filterred image in frequency domain is transferredback using 2 dimentional inverse FFT to spacial domain for visualization. The 2 dimentional FFT andinverse FFT are implemented by applying FFT first to each row, and on each column. Since FFTperforms the best on image with side length of power of 2, zero paddings are added to image to extendthe height and width of image to the nearest 2’s exponent.
See the project report for more details.