I know there is the Kernel convolution filter, as has been suggested before. But that isn't what I need. I actually need to write the filter entirely and have complete per-pixel control. I would like to prototype the filter in python, but some information on writing an extension to optimize it would be fantastic.
_______________________________________________ Image-SIG maillist - [email protected] http://mail.python.org/mailman/listinfo/image-sig
