Data filtering algorithms

Many algorithms use a mask to implement the filter. This approach only modifies the raw data and, therefore, does not offer flexibility with the amount of information we want to attenuate or enhance. A wavelet approach makes more sense, because we have more control over the frequency and time values along different resolutions.

The data filtering is carried out in three steps:

1. perform the forward wavelet transform;
2. manipulate with the spectrum;
3. perform the inverse wavelet transform.

Two ways how to manipulate with the wavelet spectrum are given below.