Information Technology Reference
In-Depth Information
(a)
⋅
1
0
−3
10
−
3
10
−
3
⋅
⋅
250
250
250
5
10
15
5
10
15
5
10
15
15.6
15.6
15.6
0.977
0.977
0.977
0.061
0.061
0.061
5
10
15
5
10
15
5
10
15
⋅
1
0
−3
10
−
3
10
−
3
⋅
⋅
250
250
250
5
10
15
5
10
15
5
10
15
15.6
15.6
15.6
(b)
0.977
0.977
0.977
0.061
0.061
0.061
5
10
15
5
10
15
5
10
15
10
−3
1
0
−
3
10
−3
⋅
⋅
⋅
250
250
250
5
10
15
5
10
15
5
10
15
15.6
15.6
15.6
0.977
0.977
0.977
0.061
0.061
0.061
5
10
15
5
10
15
5
10
15
⋅
10
−3
1
0
−3
⋅
10
−3
⋅
250
250
250
5
10
15
5
10
15
5
10
15
15.6
15.6
15.6
0.977
0.977
0.977
0.061
0.061
0.061
5
10
15
5
10
15
5
10
15
Fig. 12.6
Co-occurrence matrices
a
calculated for an example vehicle image
b
for each channel
(HSV—from
left
to
right
) and for each direction (0
ⓦ
,45
ⓦ
,90
ⓦ
, 135
ⓦ
—from
top
to
bottom
)
Five statistical parameters are calculated for every co-occurrence matrix
P
.They
were chosen from among others to be the least dependent on each other and they are
defined as follows [
13
,
25
]:
2
contrast
=
P
i
,
j
(
i
−
j
)
(12.5)
i
,
j
P
i
,
j
2
energy
=
(12.6)
i
,
j
mean
=
μ
i
=
μ
j
=
i
·
P
i
,
j
(12.7)
i
,
j
2
standard deviation
=
σ
i
=
σ
j
=
(
i
−
μ
i
)
P
i
,
j
(12.8)
i
,
j
−
μ
i
)
j
−
μ
j
P
i
,
j
(
i
correlation
=
(12.9)
σ
i
2
σ
j
2
i
,
j
This gives a total number of 60 elements for
CMSP
descriptor (3 channels
×
4
co-occurrence matrices
×
5 parameters).
Search WWH ::
Custom Search