Image Processing Reference
In-Depth Information
As the tumour/haemorrhage region is the region of interest, the image is
histogram thresholded to eliminate the unwanted pixels or clusters, thereby
keeping only the clusters required by the user. Thus, in a clustered thres-
holded image, regions that are similar to the tumour/haemorrhage pixel
are present. The thresholded image is then processed for edge detection. As
there are almost no unwanted regions, edges of the wanted region will be
present. The edge detection technique on the thresholded image will detect
the exact boundary of the tumour/haemorrhage.
Example 8.4
An example in Figures 8.7 and 8.8 shows the edges of the tumour/clot
region where the regions are extracted very clearly.
(a)
(b)
(c)
(d)
FIGURE 8.7
(a) Clot image, (b) histogram thresholded clustered image, (c) binary edge image using the intu-
itionistic fuzzy method and (d) superimposed edge image using the
IFS
method. (Modified
from Chaira, T. and Anand, S.,
J. Scient. Indust. Res.
, 70(6), 427, 2011.)
Search WWH ::
Custom Search