<s>
Pedestrian	B-General_Concept
detection	I-General_Concept
is	O
an	O
essential	O
and	O
significant	O
task	O
in	O
any	O
intelligent	B-General_Concept
video	I-General_Concept
surveillance	I-General_Concept
system	O
,	O
as	O
it	O
provides	O
the	O
fundamental	O
information	O
for	O
semantic	O
understanding	O
of	O
the	O
video	O
footages	O
.	O
</s>
<s>
Despite	O
the	O
challenges	O
,	O
pedestrian	B-General_Concept
detection	I-General_Concept
still	O
remains	O
an	O
active	O
research	O
area	O
in	O
computer	B-Application
vision	I-Application
in	O
recent	O
years	O
.	O
</s>
<s>
,	O
others	O
uses	O
local	O
features	O
like	O
histogram	B-Algorithm
of	I-Algorithm
oriented	I-Algorithm
gradients	I-Algorithm
descriptors	O
.	O
</s>
<s>
proposed	O
an	O
approach	O
combining	O
both	O
the	O
detection	O
and	O
segmentation	B-Algorithm
with	O
the	O
name	O
Implicit	O
Shape	O
Model	O
(	O
ISM	O
)	O
.	O
</s>
<s>
)	O
,	O
background	B-Algorithm
subtraction	I-Algorithm
can	O
help	O
to	O
detect	O
pedestrians	O
.	O
</s>
<s>
Background	B-Algorithm
subtraction	I-Algorithm
classifies	O
the	O
pixels	O
of	O
video	O
streams	O
as	O
either	O
background	O
,	O
where	O
no	O
motion	O
is	O
detected	O
,	O
or	O
foreground	O
,	O
where	O
motion	O
is	O
detected	O
.	O
</s>
