team-scandraid team mailing list archive

[Darius Blaszyk <dhkblaszyk@xxxxxxxxxxxxx>]

On Mon, 2009-05-04 at 23:52 +0200, arno wilhelm wrote:
> > Pre-processing steps that probably will be needed are:
> > - color to grayscale conversion
> > - histogram equalization
> > - noise reduction
> > - edge detection
> > - grayscale to bw
> > - image refinement
> >
> > Tests will be needed to determine the significance of certain
> > pre-processing steps.
> Sound good to me - think that these steps might also need methods of
> their own ...
Yes, they will. Like I said image class seems logical, as the methods 
will operate on the internal image structure.

> > Also agree on that. Do you have an idea how this could look?
> 
> Maybe we could offer a plugin interface that dynamically loads a
> library according to configuration parameters? I will have  do some
> investigation - maybe there is a operating independent library that
> can be used ....

Why not simply adopt the methods we use in the image class. Just create
a virtual VertexOutput class and override it for obj/ac3d/blender
memory?

What else comes to mind: How will we program the library itself? C or C
++ style? I would vote for C as it will be easier to be used by C type
applications and C easily mixes with C++ but not reverse (is my
understanding).

Darius