Image Processing Reference
In-Depth Information
Figure 8.16. Result of the fusion using the variable behavior
compensation operator on the Vignola scene
8.12.2. Example in medical imaging
The second example uses the same images as the example in Chapter 7 regard-
ing the classification of multi-echo MRI images of the brain (Figure 7.1). We are still
looking at the pixel level and the membership functions are defined based on the gray
levels, in an unsupervised fashion [BLO 97]. Three classes are considered: brain, ven-
tricles and cerebrospinal fluid, and the pathological area.
Here the combination operators are chosen in an adaptive fashion, not locally in
each point, but depending on the classes. Since the two images provide similar infor-
mation about the ventricles, the membership functions are fused by a mean operator.
Search WWH ::




Custom Search