<s>
In	O
digital	B-Algorithm
image	I-Algorithm
processing	I-Algorithm
,	O
thresholding	B-Algorithm
is	O
the	O
simplest	O
method	O
of	O
segmenting	B-Algorithm
images	I-Algorithm
.	O
</s>
<s>
From	O
a	O
grayscale	O
image	O
,	O
thresholding	B-Algorithm
can	O
be	O
used	O
to	O
create	O
binary	B-Algorithm
images	I-Algorithm
.	O
</s>
<s>
The	O
simplest	O
thresholding	B-Algorithm
methods	O
replace	O
each	O
pixel	O
in	O
an	O
image	O
with	O
a	O
black	O
pixel	O
if	O
the	O
image	O
intensity	O
is	O
less	O
than	O
a	O
fixed	O
value	O
called	O
the	O
threshold	O
,	O
or	O
a	O
white	O
pixel	O
if	O
the	O
pixel	O
intensity	O
is	O
greater	O
than	O
that	O
threshold	O
.	O
</s>
<s>
Many	O
types	O
of	O
automatic	O
thresholding	B-Algorithm
methods	O
exist	O
,	O
the	O
most	O
famous	O
and	O
widely	O
used	O
being	O
Otsu	B-Algorithm
's	I-Algorithm
method	I-Algorithm
.	O
</s>
<s>
(	O
2004	O
)	O
categorizes	O
thresholding	B-Algorithm
methods	O
into	O
broad	O
groups	O
based	O
on	O
the	O
information	O
the	O
algorithm	O
manipulates	O
.	O
</s>
<s>
Histogram	B-Algorithm
shape-based	O
methods	O
,	O
where	O
,	O
for	O
example	O
,	O
the	O
peaks	O
,	O
valleys	O
and	O
curvatures	O
of	O
the	O
smoothed	O
histogram	B-Algorithm
are	O
analyzed	O
.	O
</s>
<s>
Note	O
that	O
these	O
methods	O
,	O
more	O
than	O
others	O
,	O
make	O
certain	O
assumptions	O
about	O
the	O
image	O
intensity	O
probability	O
distribution	O
(	O
i.e.	O
,	O
the	O
shape	O
of	O
the	O
histogram	B-Algorithm
)	O
,	O
</s>
<s>
This	O
category	O
of	O
methods	O
is	O
called	O
local	O
or	O
adaptive	B-Algorithm
thresholding	I-Algorithm
.	O
</s>
<s>
Many	O
global	O
thresholding	B-Algorithm
methods	O
can	O
be	O
adapted	O
to	O
work	O
in	O
a	O
local	O
way	O
,	O
but	O
there	O
are	O
also	O
methods	O
developed	O
specifically	O
for	O
local	O
thresholding	B-Algorithm
,	O
such	O
as	O
the	O
Niblack	O
or	O
the	O
Bernsen	O
algorithms	O
.	O
</s>
<s>
Software	O
such	O
as	O
ImageJ	B-Language
propose	O
a	O
wide	O
range	O
of	O
automatic	O
threshold	O
methods	O
,	O
both	O
and	O
.	O
</s>
<s>
Therefore	O
,	O
the	O
HSL	O
and	O
HSV	O
color	O
models	O
are	O
more	O
often	O
used	O
;	O
note	O
that	O
since	O
hue	B-Language
is	O
a	O
circular	O
quantity	O
it	O
requires	O
circular	B-Algorithm
thresholding	I-Algorithm
.	O
</s>
<s>
Instead	O
of	O
a	O
single	O
threshold	O
resulting	O
in	O
a	O
binary	B-Algorithm
image	I-Algorithm
,	O
it	O
is	O
also	O
possible	O
to	O
introduce	O
multiple	O
increasing	O
thresholds	O
.	O
</s>
<s>
Most	O
of	O
the	O
binary	O
automatic	O
thresholding	B-Algorithm
methods	O
have	O
a	O
natural	O
extension	O
for	O
multi-thresholding	O
.	O
</s>
<s>
Thresholding	B-Algorithm
will	O
work	O
best	O
under	O
certain	O
conditions	O
:	O
</s>
<s>
In	O
difficult	O
cases	O
,	O
thresholding	B-Algorithm
will	O
likely	O
be	O
imperfect	O
and	O
yield	O
a	O
binary	B-Algorithm
image	I-Algorithm
with	O
false	O
positives	O
and	O
false	O
negatives	O
.	O
</s>
