Path: utzoo!utgpu!attcan!uunet!lll-winken!lll-lcc!ames!mailrus!cornell!rochester!udel!udccvax1!cygnus
From: cygnus@vax1.acs.udel.EDU (Marc Cygnus (ACM V.Pres.))
Newsgroups: comp.graphics
Subject: Re: Raytracing speedups....
Keywords: raytrace, octrees, bounding volumes
Message-ID: <2546@udccvax1.acs.udel.EDU>
Date: 15 Dec 88 23:47:47 GMT
References: <907@cmx.npac.syr.edu>
Reply-To: cygnus@vax1.acs.udel.EDU (Marc Cygnus (ACM V.Pres.))
Organization: The Lab Rats
Lines: 19
Organisation: The Lab Rats

In article <907@cmx.npac.syr.edu> koussari@rodan.acs.syr.edu (Vahid Koussari)
writes:
>   The following paragraphs describe an idea of mine.  This idea describe
>a possible algorithm for improving speed for raytracing...
> <etc...>
>    This process describes a possible improvement for raytracing....
>This method uses bounding boxes instead of bounding volumes to speed up
>intersection tests of objects.
> <etc, remainder deleted>

Yes, it sounds like an effective speedup. It also sounds like an octree :-).

Octree (and similar) encoding an object hierarchy to reduce computations 
is fairly common. For an introductory treatment of the idea of threespace
cell-decomposition, take a look at "A new algorithm for object oriented ray
tracing," by Saul Youssef (Computer Vision, Graphics, and Image Processing,
34, (1986), pps. 125-137). That paper proposes an interesting variation on
the octree approach. Or, look up octree encoding / 3d region decomposition.

					-marcus-