Discrete
Fourier Transform
/ Filtering
Frequency domain spectrum filtering can be performed by 2D DFT. By the transform of image
f(x,y) we obtain
F(u,v). If the H(u,v) represents filtr, filtering can then be written as:

After the forward FFT we select suitable filter and set their properties - cutoff frequency, intensity, possibly other constants.
Within the SW MIPS there are available the following filters:
We can choose whether we want to plot the real or imaginary component, the amplitude or the phase component after performing a direct FFT. At each step, the result after
the operation can be displayed in a magnified window. After the backward FFT, the resulting image can be cropped to match the size of the original image.
Preview
|

|
|
Working window for FFT filtering
|
|
|
|

|
Methods of image completion
Visualization methods
|
|
|
|

|
List of available filters
Methods of inverse DFT
|
|
|
|

|
|
Next functions
|
Examples
1.
Low-pass filter
|

|
|
Image before filtering
Image spectrum
|
|
|
|

|
|
Image spectrum after filtering Image after filtering
|
2.
High-pass filter
|

|
|
Image before filtering Image spectrum
|
|
|
|

|
|
Image spectrum after fltering Image after filtering
|
3.
Band-pass filter
|

|
|
Image
before filtering Image spectrum
|
|
|
|

|
|
Image spectrum after filtering Image after filtering
|
|
|
|