Hi

It's always nice to see new contributors. Especially, when they want to
improve the stuff I'm using :-)

> I have done a better (faster and more general) implementation
> of Euclidean and other distance transforms than what is currently
> available in the form of __bwdist() in Octave. It's unencumbered
> and free to release under the GPL.

Sounds great. Could you post the implementation to this list? Before
adding new developers we require some actual code to be posted. We do
this to keep the number of inactive developers as low as possible.

> If you add me to the list of developers, I would be glad to
> contribute the code as soon as I have read up on how to
> do proper Octave functions. Right now, it's still wrapped
> up in a MEX file, because I already knew how to do that.
> My MEX file is already about twice as fast as the existing bwdist(),
> it can be extended to support more than the Euclidean distance,
> and it can return the full array of output values which the Matlab
> version of the function does, i.e. you can get not only the distance
> but also the index of the closest feature pixel.

This sounds really nice. Is it a big problem to port the function from
MEX to using Octave's functions? For maintenance reasons I'd like to
avoid having MEX files in the 'image' package.

Søren


------------------------------------------------------------------------------
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev

Reply via email to