This link contains the following statement:

In contrast to the frequency domain, it is possible to implement non-linear filters in the spatial domain. In this case, the summations in the convolution function are replaced with some kind of non-linear operator.

Why is it not possible to perform non-linear filtering in frequency domain ?

  • $\begingroup$ You could, by adding spectral components in just the right amounts, but it would be far more complicated than just doing it in the time domain. Likewise, non-linear frequency processing is much easier to do in the frequency domain than the time domain. $\endgroup$ – endolith May 6 '13 at 23:57

Linear filters in the spatial domain have a direct equivalent in the frequency domain, so you can transform your data and filter between spatial<->frequency domains and get equivalent results. However the transforms that we use for converting between spatial<->frequency domains are only valid for linear systems - a non-linear filter such as a median filter therefore has no frequency domain equivalent, so we can't perform e.g. median filtering in the frequency domain.

Note that this does not mean that there are no non-linear operations that we can perform in the frequency domain, e.g. we could reduce all frequency components below a certain threshold value to zero, it's just that there are no counterparts between the two domains for non-linear operations, so the frequency domain threshold operation described would have no equivalent operation in the spatial domain.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.