Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!purdue!lerc.nasa.gov!magnus.acs.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!swrinde!sgigate.sgi.com!genmagic!bug.rahul.net!a2i!olivea!charnel.ecst.csuchico.edu!csusac!csus.edu!netcom.com!whitten
From: whitten@netcom.com (David Whitten)
Subject: Re: Modelling human anatomy
Message-ID: <whittenDCGM95.Gr0@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
X-Newsreader: TIN [version 1.2 PL1]
References: <3uip5v$2cv@sun0.urz.uni-heidelberg.de>
Date: Sat, 29 Jul 1995 04:10:17 GMT
Lines: 37
Sender: whitten@netcom11.netcom.com

Johannes Link (johannes@borneo.dkfz-heidelberg.de) wrote:
: I'm doing research in the field of knowledge-based interpretation of medical
: images and I'm currently working on combining knowledge about medical 
: imaging and human anatomy.
: Are there any known (generic) models of human anatomy - or parts of it -,
: which describe more than just generalisation and aggregation hierarchies,
: but regard as well things like topological and geometric relations between
: anatomic parts and their (3D-)shape?
: Any hints are appreciated. Thanx in advance.
: J.Link

I've been thinking about this, and I think it is really hard.
most of the methods I have heard of expressing knowledge are inherently
one dimensional.

If you want to express the relations between the various parts of a
'typical human head', I have no idea what the best representation would be.

I can think of listing parts as nodes, and then providing 'is-near-to' links
to link all the parts together,  but this really doesn't express teh fact
that a cheek may cover more surface area of the head than a nostril.

I can list parts in terms of a deformed sphere, and specify the longitude
and latitude of each part, but this doesn't give you any idea of the shape
and barely an idea of what parts are 'near' each other.
but of course each part is an area, not a point.

you can list all parts that are on the same horizontal slice across the
face, but this separates the parts that are on the same vertical slice of
the face. 


I know some details of Conceptual graph logics, but they are a logic that
uses a 'page of assertions' but not necessary a logic about two dimensional
or three dimensional figures.

David (whitten@netcom.com) (214) 437-5255
