home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!wupost!howland.reston.ans.net!paladin.american.edu!gatech!usenet.ins.cwru.edu!agate!ames!nsisrv!mimsy!afterlife!rmrohre
- From: rmrohre@afterlife.ncsc.mil (Randile the Reptile)
- Newsgroups: comp.graphics
- Subject: Re: Fast front->back hsr for voxels
- Message-ID: <1993Jan4.180930.27923@afterlife.ncsc.mil>
- Date: 4 Jan 93 18:09:30 GMT
- References: <38127@uflorida.cis.ufl.edu>
- Organization: The Great Beyond
- Lines: 26
-
- In article <38127@uflorida.cis.ufl.edu> LIONESS@oak.circa.ufl.edu writes:
- >
- >Have there been any papers for fast hidden surface removal using a
- >front to back method with voxels? A painter's algorithm would work,
- >but on systems with slow video memory access ( such as a VGA equipped
- >PC w/o local bus ) you would spend most of your time drawing things
- >that would later be overwritten. Given that you are dealing with,
- >in effect, point clouds, is there an efficient way of determining what
- >part of a point cloud is visible and what is not?
-
- I suspect that you're looking for "object precision"
- approaches (since you asked about front-to-back methods).
- However, Marc Levoy's ray casting method ("image precision")
- terminates rays passing through voxel space as soon as full opacity
- has been accumulated for each ray. This speeds up rendering
- (and "hiden voxel removal") by avoiding processing voxels that do
- not contribute to the current pixel in the image.
-
- Randy
-
-
-
- --
- Randy Rohrer && The George Washington University
- rmrohre@afterlife.ncsc.mil Depratment of EE&CS
- rohrer@seas.gwu.edu
-