Image Processing Reference
In-Depth Information
E x a mple 7. 3
An example of a brain clot/haemorrhage image is shown. It is observed
that the clustered image without kernel using intuitionistic fuzzy set
does not cluster image clearly as it contains noise, but it is better than
fuzzy clustering (Figure 7.5).
2. Chaira [3,4] suggested another intuitionistic clustering algorithm. In the
construction of the algorithm, the following steps are required: (a) creat-
ing an intuitionistic fuzzy set using Yager-type intuitionistic fuzzy gen-
erator and (b) modifying the objective criterion function, updating the
cluster centre using intuitionistic fuzzy set and incorporating an objec-
tive function, that is, the intuitionistic fuzzy entropy (
IFE
), in the criterion
function.
An objective function, that is,
IFE
, is introduced in the conventional clus-
tering algorithm. The
IFE
by Chaira [4] is given as (described in Chapter 4)
n
∑
π
()
[
1
−
π
Ai
(
)]
IFEA
()
=
xe
Ai
⋅
( 7. 2 0 )
i
=
1
where π
A
(
x
i
) = 1
−
(μ
A
(
x
i
) + ν
A
(
x
i
)).
(a)
(b)
FIGURE 7.5
(a) CT scan brain image and (b) intuitionistic fuzzy method by Dimitris.
Search WWH ::
Custom Search