[Help-gsl] Spatial filtering of an image using wavelet transforms

[Sebastien Maret]

Hi,

I would like to write a program to filter-out the large scale spatial
variations from an image, using a wavelet transform. For this I have
padded the image with zeros so it's dimensions are 2*n * 2*n, and I have
computed the wavelet transform. I assume that in order to filter-out the
large variation scales (e.g. larger than 10 pixels), I need to set some
of these coefficients to zero and then do the backward
transform. However I am not sure about which coefficients I should to be
set to zero in order to do the filtering. Has anybody done that before?

A couple of side questions: what kind of wavelet should I choose for the
transform? Is it okay to pad the image with zeros, or is it better to
make the image periodic?

Thanks in advance for your help,
Sebastien