Image Processing Reference
In-Depth Information
Similarity Matrix − Sim1 − TOD
Similarity Matrix − Sim2 − tNM
1
1
airplane1.jpg−[1]
airplane1.jpg−[1]
0.9
0.9
airplane2.jpg−[2]
airplane2.jpg−[2]
0.8
0.8
airplane3.jpg−[3]
airplane3.jpg−[3]
0.7
0.7
face_0009.jpg−[4]
0.6
face_0009.jpg−[4]
0.6
face_0073.jpg−[5]
0.5
face_0073.jpg−[5]
0.5
0.4
face_0095.jpg−[6]
0.4
face_0095.jpg−[6]
image_0008.jpg−[7]
0.3
0.3
image_0008.jpg−[7]
0.2
0.2
image_0011.jpg−[8]
image_0011.jpg−[8]
0.1
0.1
image_0021.jpg−[9]
image_0021.jpg−[9]
0
0
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9
FIGURE 8.15: TOD and tNM similarity matrix for images in 8.14
database (Computational Vision Group at Caltech, 2009). The set consists of 3
subsets of airplanes, faces and leaves images. The similarity matrices for TOD and
tNM are shown in figure 8.15 where
B
=
{φ
,φ
}
, the set of probe functions consists
1
2
of
φ
2
(entropy) of subimages. Feature values have
been normalized between 0 and 1 and
φ
1
(the average gray level) and
ε
=0
.
1.
FIGURE 8.16: Selected images from USC-SIPI dataset
Example 8.9
As an example, 50 images in 5 different groups are randomly selected from USC-
SIPI image database (USC Signal and image processing institute, 2009) shown in
figure 8.16 where each group of images is shown in one row. TOD and tNM are
Search WWH ::
Custom Search