Image Processing Reference
In-Depth Information
)
(
i
i
i
i
y
=
ψ
x
,
x
,
x
,
(8.10)
01
T
−
1
)
(
This means that a stack filter need only compute as many threshold levels as
there are unique values in the input window. So a stack filter with
T
input variables
may be implemented by working with, at most,
T
threshold levels rather than com-
pute the full dynamic range, typically 256 for 8-bit data. This method is called
“range compression” by Lin et al.
19
Range compression is illustrated in Fig. 8.8.
=
where
x
∈
x
,
x
,
x
.
then
yx
j
j
01
T
−
1
Figure 8.8
Range compression techniques for stack filter implementation. (a) Identification
of ranges. (b) Compression of ranges to only those that may potentially form the output value.