Thresholding of Medical Images - Medical Image Processing: Advanced Fuzzy Set Theoretic Techniques - page 125

(a) (b) (c)

FIGURE 6.4
(a) Blood vessel image, (b) intuitionistic fuzzy method by Vlachos and (c) intuitionistic fuzzy
method by Chaira.

For all threshold grey levels, the intuitionistic fuzzy divergence is
calculated. The threshold level corresponding to the minimum divergence
is selected as the optimal threshold.

Example 6.4

Two examples of medical images (by Vlachos and Chaira) are shown
in Figure 6.4 to illustrate the efficacy of intuitionistic fuzzy divergence
thresholding methods. Figure 6.4b is the result using Vlachos' method,
and Figure 6.4c is the result using Chaira's method.

6.5 Window-Based Thresholding
Local thresholding is useful in medical images as it selects the thresh-
old for each window. As medical images are poorly illuminated, global
threshold may not work better. In pathological images, where the blood
vessels are hardly visible, window-based thresholding will work better. In
window-based thresholding, the threshold is selected for each window, so
the image will be thresholded depending on the image characteristics for
t h at w i ndow.
Chaira [8] suggested a window-based method for image thresholding. In
this method, the restricted equivalence function is used to find the member-
ship function.