<%@ include file="../../check-login.jsp" %> Subdivision.org: Re: bounding boxes for loop's subdivision patches <% session.setAttribute ( "messageNum", new Integer ( 2617 ) ); %>

Re: bounding boxes for loop's subdivision patches

Posted by Scott Schaefer on September 15, 2005 at 10:18:11:

In Reply to: Re: bounding boxes for loop's subdivision patches posted by Ray on September 14, 2005 at 13:44:15:

: Thanks Scott! I am curious if there is any tighter estimation than this. Take a cube for example, the bounding box for any base patch is actually the whole cube, which seems too conservative. Any thoughts?

Well, finding tighter bounds is more problematic. The typical way is to subdivide your triangle patch into four sub-patches. Take the convex hull of all of the bounding hulls of the sub-patches.

In general it's very difficult to deal with extraordinary vertices. Jorg Peters has done a lot of work about computing tight enclosures for various spline surfaces. However, all of his papers either deal with curves or polynomial surfaces. Unfortunately, subdivision surfaces do not fall within this class of shapes.

Follow Ups:

Post a Followup

Name:
E-Mail:

Subject:

Comments:
: : Thanks Scott! I am curious if there is any tighter estimation than this. Take a cube for example, the bounding box for any base patch is actually the whole cube, which seems too conservative. Any thoughts? : Well, finding tighter bounds is more problematic. The typical way is to subdivide your triangle patch into four sub-patches. Take the convex hull of all of the bounding hulls of the sub-patches. : In general it's very difficult to deal with extraordinary vertices. Jorg Peters has done a lot of work about computing tight enclosures for various spline surfaces. However, all of his papers either deal with curves or polynomial surfaces. Unfortunately, subdivision surfaces do not fall within this class of shapes.

Optional Link URL:
Link Title:
Optional Image URL: