Meta-learning for Image Processing Based on Case-Based Reasoning - Computational Intelligence in Healthcare: Advanced Methodologies - page 232

(a) cell (b) gan128 (c) neu1 (d) neu2 (e) neu3

(f) neu4 (h) neu4_r180 (i) monroe (j) parrot

Fig. 16. Images used for the study

Table 2. Segmentation parameters a, b, and T of the test images shown in Fig. 16

ImageName a b T

monroe 0,5 2 1

gan128 0,75 0,75 1

parrot 0,75 2 1

cellr 1 1 1

neu1 2 1 1

neu2 0,25 1 1

neu3 2 1 1

neu4 0,75 2 1

neu4_r180 1 0,5 1

Table 3 shows the Images having the same segmentation parameters. Based on
segmentation parameters the images neu1 and neu3 and neu4 and parrot should be
grouped together by the image description.

Table 3. Images having the same segmentation parameters

ImageName a b T

parrot 0,75 2 1

neu4 0,75 2 1

neu1 2 1 1

neu3 2 1 1

Fig. 17 and Table 3 demonstrate the clustering of images grouped according to
their parameters. Based on this result we found an image description which grouped
the images monroe, parrot and neu4 as well as the images neu1 and neu3.