Digital Signal Processing Reference
In-Depth Information
(a)
(b)
(c)
(d)
Figure 11.7
Quantitative analysis of clustering results with regard to side asymmetry of brain
perfusion in analogy to figure 11.4 for fuzzy
c
-means vector quantization. For a
better orientation, an anatomic EPI scan of the analyzed slice is underlaid in (a)
and (b). The X-axis represents the scan number, and the Y-axis is arbitrary for (c)
and (d).
tering methods in automatically evolving the appropriate number of
clusters is demonstrated experimentally in the form of cluster assign-
ment maps for the perfusion MRI data sets, with the number of clusters
varying from 2 to 36.
Table 11.2 shows the optimal cluster number
K
∗
obtained for each
perfusion MRI data set, based on the different cluster validity indices.
Figures 11.11 and 11.12 show results for cluster-validity analysis for
Search WWH ::
Custom Search