home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!destroyer!ubc-cs!unixg.ubc.ca!unixg.ubc.ca!israel
- From: israel@unixg.ubc.ca (Robert B. Israel)
- Subject: Re: Mapping points in space to a plane
- Message-ID: <israel.714343659@unixg.ubc.ca>
- Sender: news@unixg.ubc.ca (Usenet News Maintenance)
- Nntp-Posting-Host: unixg.ubc.ca
- Organization: University of British Columbia, Vancouver, B.C., Canada
- References: <BtAIpG.5pL@wpi.WPI.EDU>
- Distribution: usa
- Date: Thu, 20 Aug 1992 20:47:39 GMT
- Lines: 59
-
- In <BtAIpG.5pL@wpi.WPI.EDU> jac@cs.WPI.EDU (Jeffrey A Choate) writes:
-
-
- >Hello,
-
- >I am having a problem mapping a point to a plane,
- >which I hope someone can help with.
-
- >The problem is as follows:
-
- >I have a point which lies on a plane with the
- >equation: x + y + z = 1 (point represents
- >probabilities). After I perform some non-linear
- >transformation, the point no longer necessarily lies
- >on the "probabilistic" plane.
-
- >I have to map this resulting point, which lies
- >somewhere in 3-space back to the probabilistic
- >plane.
-
- >My question is how to best map this point back to
- >the plane?
-
- >One suggestion is to find the orthogonal projection
- >of the point on the plane. Another suggestion is to
- >use the normal to the plane (1,1,1), and starting at
- >the point follow the normal vector back until it
- >intersects the plane, and use the intersection point.
-
- >Are any of these ideas good? Is there an algorithmic
- >way of doing this? (I have to perform this mapping for
- >every point in a very large data set) Any help would be
- >greatly appreciated.
-
- >Thank you,
-
- >Jeff Choate
-
- >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- >~ Jeff Choate | (508) 831-5006 | Computer Science Department ~
- >~ jchoate@cs.wpi.edu | AI Research Group | Worcester Polytechnic Institute ~
- >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
- You're in a Research Group at a Computer Science department and you don't
- know first-year Linear Algebra? What's this world coming to?
-
- Both suggestions amount to the same thing. The formula for the
- projection is P v + b where v is your original point (as a column vector),
- P is the matrix
- [ 2/3 -1/3 -1/3 ]
- [ -1/3 2/3 -1/3 ]
- [ -1/3 -1/3 2/3 ]
- and b is the vector (1/3, 1/3, 1/3) (transposed).
-
- --
- Robert Israel israel@math.ubc.ca
- Department of Mathematics or israel@unixg.ubc.ca
- University of British Columbia
- Vancouver, BC, Canada V6T 1Y4
-