Didier

Thanks for your reply.

I agree these are two different approaches. But the idea that this technology will find its way into consumer electronics stuck in my head. I think future mobile devices could do the job, irrepsective of high polygon counts. Recording a 3D scene with a move of my hand would be awesome.

Even if user assisted processing has its advantages I would say that in the long term machines should be using this and will be able to truly recognize things. I mean, humans do it

123D Catch isn’t quite flawless, but if you keep it simple you can achieve amazing results. Also, with this, a simple camera can replace a 3D laser scanner or a similar system.

Staying tuned & best regards,
Lernsky

I think the Autodesk 123D gear is great, but I suggest you really give it a go before you get too enthusiastic about it.

It generates high polygon count models of only what you can see, which is very different to the Videotrace approach where you get only the polygons you want, even of bits of the object that you can’t see in the image set.

They are different approaches for different purposes.

It seems like Autodesk has done it:

http://www.123dapp.com/catch

Enjoy!

We’d have to find a way to estimate the camera parameters for medical scans. I’m sure the scanners provide that information, but we’d need to get it into Videotrace in a form we could use.

Let us know if you have any great ideas.