Biomedical Engineering Reference
In-Depth Information
b
E
(
Y
)=0
E
(
Y
)=1
E
(
Y
)=2
E
(
Y
)=3
U1
U1
U1
U1
U2
U2
U2
U2
U3
U3
U3
U3
U4
U4
U4
U4
U5
U5
U5
U5
U6
U6
U6
U6
U7
U7
U7
U7
U8
U8
U8
U8
U9
U9
U9
U9
Fig. 4.7b Permutation pattern for
Y
coordinate
To code the feature
F
s
location in the image, it is necessary to select the
correlation distance
D
c
. Let the same feature
F
s
be detected in two different points,
P
1
and
P
2
. If the distance
d
between them is
<
D
c
, the corresponding codes will be
correlated. If the distance
d
D
c
, the codes will be uncorrelated. To obtain this
property, we have to calculate the following values:
>
X
¼
j
=
Dc
;
(4.6)
EX
ðÞ¼
ðÞ
int
X
;
RX
ðÞ¼
j
EX
ðÞ
Dc
;
Y
¼
i
=
Dc
;
(4.7)
EY
ðÞ¼
ðÞ
int
Y
;
RY
ðÞ¼
i
EY
ðÞ
Dc
;
(4.8)
R
ð
X
Þ
N
Px
¼
ðÞ
int
D
c
R
ð
Y
Þ
N
Py
¼
ðÞ
int
;
(4.9)
D
c
where
E
(
X
) is the integer part of
X
;
R
(
X
) is the fraction part of
X
;
i
is the vertical
coordinate of the detected feature;
j
is the horizontal coordinate of the detected
feature,
N
is the number of neurons; and
P
x
,
P
y
are the fractions of the neurons for
which an additional permutation is needed.