<s>
In	O
the	O
fields	O
of	O
computing	O
and	O
computer	B-Application
vision	I-Application
,	O
pose	B-Architecture
(	O
or	O
spatial	O
pose	B-Architecture
)	O
represents	O
the	O
position	O
and	O
orientation	O
of	O
an	O
object	O
,	O
usually	O
in	O
three	O
dimensions	O
.	O
</s>
<s>
Poses	B-Architecture
are	O
often	O
stored	O
internally	O
as	O
transformation	B-Algorithm
matrices	I-Algorithm
.	O
</s>
<s>
The	O
term	O
“	O
pose	B-Architecture
”	O
is	O
largely	O
synonymous	O
with	O
the	O
term	O
“	O
transform	O
”	O
,	O
but	O
a	O
transform	O
may	O
often	O
include	O
scale	B-Algorithm
,	O
whereas	O
pose	B-Architecture
does	O
not	O
.	O
</s>
<s>
In	O
computer	B-Application
vision	I-Application
,	O
the	O
pose	B-Architecture
of	O
an	O
object	O
is	O
often	O
estimated	O
from	O
camera	O
input	O
by	O
the	O
process	O
of	O
pose	B-General_Concept
estimation	I-General_Concept
.	O
</s>
<s>
The	O
specific	O
task	O
of	O
determining	O
the	O
pose	B-Architecture
of	O
an	O
object	O
in	O
an	O
image	O
(	O
or	O
stereo	O
images	O
,	O
image	O
sequence	O
)	O
is	O
referred	O
to	O
as	O
pose	B-General_Concept
estimation	I-General_Concept
.	O
</s>
<s>
Pose	B-General_Concept
estimation	I-General_Concept
problems	O
can	O
be	O
solved	O
in	O
different	O
ways	O
depending	O
on	O
the	O
image	O
sensor	O
configuration	O
,	O
and	O
choice	O
of	O
methodology	O
.	O
</s>
<s>
If	O
also	O
the	O
geometry	O
of	O
the	O
object	O
is	O
known	O
,	O
it	O
means	O
that	O
the	O
projected	O
image	O
of	O
the	O
object	O
on	O
the	O
camera	O
image	O
is	O
a	O
well-known	O
function	O
of	O
the	O
object	O
's	O
pose	B-Architecture
.	O
</s>
<s>
Once	O
a	O
set	O
of	O
control	O
points	O
on	O
the	O
object	O
,	O
typically	O
corners	O
or	O
other	O
feature	O
points	O
,	O
has	O
been	O
identified	O
,	O
it	O
is	O
then	O
possible	O
to	O
solve	O
the	O
pose	B-Architecture
transformation	O
from	O
a	O
set	O
of	O
equations	O
which	O
relate	O
the	O
3D	O
coordinates	O
of	O
the	O
points	O
with	O
their	O
2D	O
image	O
coordinates	O
.	O
</s>
<s>
Algorithms	O
that	O
determine	O
the	O
pose	B-Architecture
of	O
a	O
point	B-Algorithm
cloud	I-Algorithm
with	O
respect	O
to	O
another	O
point	B-Algorithm
cloud	I-Algorithm
are	O
known	O
as	O
point	B-Algorithm
set	I-Algorithm
registration	I-Algorithm
algorithms	O
,	O
if	O
the	O
correspondences	O
between	O
points	O
are	O
not	O
already	O
known	O
.	O
</s>
<s>
Genetic	B-Algorithm
algorithm	I-Algorithm
methods	O
:	O
If	O
the	O
pose	B-Architecture
of	O
an	O
object	O
does	O
not	O
have	O
to	O
be	O
computed	O
in	O
real-time	O
a	O
genetic	B-Algorithm
algorithm	I-Algorithm
may	O
be	O
used	O
.	O
</s>
<s>
In	O
this	O
particular	O
case	O
,	O
the	O
pose	B-Architecture
represent	O
the	O
genetic	B-Algorithm
representation	I-Algorithm
and	O
the	O
error	O
between	O
the	O
projection	O
of	O
the	O
object	O
control	O
points	O
with	O
the	O
image	O
is	O
the	O
fitness	O
function	O
.	O
</s>
<s>
Learning-based	O
methods	O
:	O
These	O
methods	O
use	O
artificial	O
learning-based	O
system	O
which	O
learn	O
the	O
mapping	O
from	O
2D	O
image	O
features	O
to	O
pose	B-Architecture
transformation	O
.	O
</s>
<s>
In	O
short	O
,	O
this	O
means	O
that	O
a	O
sufficiently	O
large	O
set	O
of	O
images	O
of	O
the	O
object	O
,	O
in	O
different	O
poses	B-Architecture
,	O
must	O
be	O
presented	O
to	O
the	O
system	O
during	O
a	O
learning	O
phase	O
.	O
</s>
<s>
Once	O
the	O
learning	O
phase	O
is	O
completed	O
,	O
the	O
system	O
should	O
be	O
able	O
to	O
present	O
an	O
estimate	O
of	O
the	O
object	O
's	O
pose	B-Architecture
given	O
an	O
image	O
of	O
the	O
object	O
.	O
</s>
