[ptx] autopano #2, using SIFT

Pablo d'Angelo pablo at mathematik.uni-ulm.de
Fri Jan 16 20:14:32 GMT 2004


Hi Alexandre,

> My opinion is that .NET is really cool for gui and interfacing beetween
> Languages or anything like that.

I'm sure .NET is a lot better than wxWindows for that (from a programmers
perspective).

> But for calculation, we should definitely
> Keep C/C++. So this hugin project generator will be in this language.
> Of course, using the same codebase is the good option.

ack. I have started to separate some of the image processing/calculation
stuff that has sneaked into the applications.

> > Something I'd like to see in any case is a clone of the 
> > "keypoints" program, available on Dr. Lowe's site :)
> 
> Keypoints just gives keypoints : it does not the matching.

Jep, thats why it is easy to call it externally, and read in the features.
even if the rest is written in c++, or available only as binary. I'm
currently researching with Dr. Lowe's linux binary implementation, but I
won't have time to learn and implement really advanced matching algorithms,
so I leave that to you.

> N-space neightboor finding sheme to hough transform is free. So if any issue
> arise from the using of sift feature, we could easy switch to a haris
> coner feature or Anyelse with keeping the matching sheme and matrices
> calculus.

*g* jep. Hmm, we should think about a code structure for that, so that there
is a clean interface to the matching, whatever algorithm is used.

ciao
  Pablo
-- 
http://wurm.wohnheim.uni-ulm.de/~redman/
Please use PGP


More information about the ptX mailing list