| View previous topic :: View next topic |
| Author |
Message |
gziegler
.
Joined: 05 May 2004
Posts: 21
Location: Saarbrücken
|
 Posted: Fri Oct 05, 2007 12:16 pm Post subject: HistoPyramids for GPU Data Expansion |
 |
|
Hej !
You might have seen our Eikonal Rendering simulator, ljus, on SIGGRAPH - it uses HistoPyramids to re-organize the wavefront while it passes through the object. We can both eliminate and tesselate wavefront patches, and are thus able to re-produce geometry shader functionality 
Here is a teaser image:
You find the principles explained in the following slides:
http://www.mpi-inf.mpg.de/~gziegler/ljus/eikosketch_presentation.pdf
And I made the code for our light simulator available on http://nvision.sf.net
Questions and comments welcome
_________________
/Gernot |
|
| Back to top |
|
 |
Ferdi
.
Joined: 25 Apr 2007
Posts: 17
Location: Amsterdam
|
| Posted: Tue Oct 09, 2007 12:36 pm Post subject: |
|
|
| Interesting, I'll read through it. I actually recently cited your other paper on the histo-pyramids in a publication as a possible optimization to 2.5D image warping. I didn't implement it though; I should still try it in the future. |
|
| Back to top |
|
|
gziegler
.
Joined: 05 May 2004
Posts: 21
Location: Saarbrücken
|
| Posted: Tue Oct 09, 2007 12:44 pm Post subject: Use in 2.5D image warping |
|
|
Ah ! Please mail me one time, maybe I can assist in how it can be applied
_________________
/Gernot |
|
| Back to top |
|
|
Ferdi
.
Joined: 25 Apr 2007
Posts: 17
Location: Amsterdam
|
| Posted: Wed Oct 17, 2007 8:54 am Post subject: Re: Use in 2.5D image warping |
|
|
| gziegler wrote: | | Ah ! Please mail me one time, maybe I can assist in how it can be applied |
Sure, we really need to get it faster, so... anyway I'll be presenting it at vrst next month. Not sure if I'm allowed to send the paper around yet. |
|
| Back to top |
|
|
gziegler
.
Joined: 05 May 2004
Posts: 21
Location: Saarbrücken
|
| Posted: Sun Dec 23, 2007 9:38 pm Post subject: |
|
|
To show that we now are able to replace geometry shader functionality, we have recently published a tech report on HistoPyramid based Marching Cube.
And, believe or not: we are currently a lot faster than geometry shader based approaches, and even CUDA 
Find more info on my research homepage http://www.mpii.de/~gziegler, it also contains two YouTube videos for your enjoyment
_________________
/Gernot |
|
| Back to top |
|
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
