Image Processing Reference
In-Depth Information
(a)
(b)
FIGURE 7.4
(a) CT scan brain image and (b) kernel clustered image.
7.6 Intuitionistic Fuzzy
c
Means Clustering
Uncertainty is largely present in medical images because of noise in image
acquisition and low illumination. This is the reason why the regions or
boundaries are not clear. Fuzzy set theory works well on these images as
it considers imprecise information. But the intuitionistic fuzzy set theory
considers more information compared to fuzzy set. It considers both the
membership and non-membership degrees. As it considers more uncertain-
ties, intuitionistic fuzzy set is assumed to work better on medical images.
Medical images are very difficult to analyse, and there is very little work
on clustering using intuitionistic fuzzy set on medical images. It follows a
similar type of algorithm as FCM but in an intuitionistic way. To see how the
intuitionistic property is incorporated into clustering, some works suggested
by different authors are described in this section:
1. Iakovidis et al. [9] suggested a new similarity metric in conventional clus-
tering algorithm where both membership and non-membership degrees
are included. The similarity metric is based on the histogram intersection
technique. Let
A
(μ
A
,
v
A
) and
B
(μ
B
,
v
B
) be two intuitionistic fuzzy sets where
Search WWH ::
Custom Search