COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	{λ30FAαINTRODUCTION.P1I425,0JCFA}   SECTION 0.
C00007 00003	{W0,675JUFA
C00011 ENDMK
C⊗;
{λ30;FA;αINTRODUCTION.;P1;I425,0;JCFA}   SECTION 0.
{JCFD}                      INTRODUCTION.
	
{JU;λ7;FM}
	"For  the purpose  of  presenting my  argument  I must  first
explain the basic  premise of sorcery as don Juan presented it to me.
He said that for a sorcerer, the world of everyday life is  not real,
or out  there, as we  believe it is. For  a sorcerer, reality  or the
world we  all know, is only a description. For the sake of validating
this premise  don  Juan concentrated  the best  of  his efforts  into
leading me  to a genuine conviction  that what I held in  mind as the
world at hand was  merely a description of  the world; a  description
that had been pounded into me from the moment I was born."

{JR;FM} - Carlos Castaneda. Journey to Ixtlan.


{JU;I1150,0;λ30;FA}
	This thesis  is about  computer techniques  for handling  3-D
geometric descriptions  of the world; the world  that can be visually
perceived with a television camera.   The overall design idea may  be
characterized as  an inverse  computer graphics approach  to computer
vision. In  computer graphics, the world is represented in sufficient
detail so that the image forming process can be numerically simulated
to  generate synthetic television  images; in the  inverse, perceived
television pictures (from a real  TV camera) are analysed to  compute
detailed geometric models. For example,  the  polyhedra in figure 0.1
on page two were computed from views of a plastic horse on a turntable.
It is hoped,  that visually  acquired 3-D geometric models can be  of
use to other  robotic processes such as manipulation,   navigation or
recognition.{Q;
L0,475;H2;*HORS01.PLT;
L0,-475;H2;*HORS02.PLT;
L0,0;JA;JC;FA} FIGURE 0.1  -  HORSE SHAPED POLYHEDRA DERIVED FROM VIDEO IMAGES.
{W0,675;JU;FA;
}	Once acquired,   a 3-D  model can be  used to  anticipate the
appearance  of an object in  a scene, making  feasible a quantitative
form of visual  feedback.  For  example,  the  appearance of the  two
machine parts depicted  in figure 0.2 can be  computed and anaylsized
(top half  of figures  0.3 and 0.4) and  compared with an  anaylsis of an actual
video image of  the parts  (bottom half of
figures 0.3 and  0.4). By comparing  the
predicted image  with a perceived image,   the correspondence between
features of the internal model  and features of the external  reality
can be  established and  a corrected  location of  the parts and  the
camera can be measured.
{W0,1250;↓;I300,800;FA}FIGURE 0.2{H2;L415,510;*PUMP02;↑;JU;FA;}
	Finally by way of introduction, I wish  to emphasive that the
kind  of vision being  attempted is  metric rather than  semantic and
that the results achieved to date are modest.  Feature classification
and recognition  in terms  of English words  is not  being attempted,
rather  a system  of prediction  and correction  between a  3-D world
model and a sequence of images is contemplated. The  chapters of this
thesis proceed twice from  theory through an implementation, with the
first five chapters dealing with modeling and the last five  chapters
dealing with vision. Theory  on geometric modeling is in  chapter one
and  theory on computer  vision in  chapter six.   The implementation
consists of  two main  programs named  GEOMED and  CRE.  GEOMED is  a
system of 3-D modeling routines with which arbitrary polyhedra may be
constructed,  altered,   or viewed  in perspective with  hidden lines
eliminated; and CRE is a solution to the problem of finding intensity
contours  in  a  sequence  of  television  pictures  and  of  linking
corresponding contours between pictures.
{Q;L0,475;*PUMP3.VID;L0,-475;*PMP3.VID;
L0,0;JA;JC;FA} FIGURE 0.3  -  PREDICTED VIDEO ↑ AND PERCEIVED VIDEO ↓.
{Q;L0,475;H2;*PUMP04.PLT;L0,-475;H2;*PUMP03.PLT;
L0,0;JA;JC;FA} FIGURE 0.4  -  PREDICTED IMAGE ↑ AND PERCEIVED IMAGE ↓.