Image Processing Reference
In-Depth Information
N
u
1
,
n
1
N
u
1
Cp
u
1
,
n
1
¼
ð
12
:
4
Þ
where
N
u
1
,
n
1
is the number of times that
u
1
has been connected to the 3D immersive
communications framework via node
n
1
.
Aging factors and Markov chains could
also be introduced, but would further complicate the approach. A simplified
Location Connectivity Variance
LCVp
(
u
1
) of user
u
1
could be defined as:
t
X
Nn
i¼
1
Cp
u
1
,
i
2
ð
Cp
u
1
Þ
LCVp
u
1
¼
1
N
n
ð
12
:
5
Þ
N
n
1
where
N
n
is the number of different geographical nodes/locations that user
u
1
has
been connected to the social network and
Cp
u
1
is the average probability of
u
1
to be
connected via any node/location. Equation (12.5) can also be written as:
p
N
n
LCVp
u
1
¼
1
s
u
1
ð
12
:
6
Þ
Where
s
u
1
is the corrected sample standard deviation of the average probability of
u
1
to be connected via any node/location. As a result, 1
s
u
p
is the
probability that
u
1
remains at the same network location (the same geographical
location/node).
Based on the above, we define as Location Dependent Interaction probability
LDIp
u
1
,
u
2
N
n
LCVp
u
1
¼
the probability of
u
1
to interact with
u
2
, having both
u
1
and
u
2
static as:
LDIp
u
1
,
u
2
Ip
u
1
,
u
2
LCVp
u
1
a
LCVp
u
2
a
¼
ð
1
Þ
ð
1
Þ
ð
12
:
7
Þ
where
1 that shows the importance that we would like to give
to the location connectivity pattern of the users. The
LDIp
u
1
,
u
2
can be used for the
overlay hierarchy optimization, as it directly associates the users
u
1
and
u
2
.
is a small integer
>
α
12.5 Node Selection Based on Social Interaction
Taking into account the social network interaction analysis and connectivity vari-
ation pattern of users, we may select a node
A
as having an optimal location to serve
as overlay node with respect to node
B
based on the following algorithm:
• Step 1: Based on Eq. (12.4), select a group
U
1
with the
k
1
users that most often
connect to the overlay using node
B
as access node.
