Image Processing Reference
In-Depth Information
20
11
Mean
95%
Mean
95%
10
18
9
16
8
14
7
12
6
10
5
8
4
6
10
1
10
2
10
3
10
4
10
1
10
2
10
3
10
4
No. of samples
No. of samples
FIGURE 7.18
The accuracy of Neugebauer model plotted as a function of the number of
characterization samples.
Evidently, the accuracy of Neugebauer models cannot be improved beyond a
certain number based on increased training samples. Since nonlinearities of the
physical process are captured in the function, the model gives reasonably good
prediction within a few hundred training samples (Table 7.5).
TABLE 7.5
Accuracy of Neugebauer Models (
D
E
2000
and
D
E
a
*
Space)
D
E
2000
Max Color
No. of Samples
Mean
95%
Max
L*
a*
b*
81
4.66
10.03
15.14
18.76
5.30
28.86
256
4.60
10.27
16.39
9.29
17.24
0.23
625
4.24
9.45
17.51
7.87
18.33
1.82
1296
4.19
9.35
17.86
6.98
18.34
2.30
2401
4.25
9.19
17.79
7.33
18.44
2.36
4096
4.11
9.61
18.03
4.60
16.62
3.14
6561
4.10
9.51
18.07
4.98
17.07
2.96
10000
4.17
9.24
18.14
5.28
17.50
2.82
D
E
a
*
Max Color
No. of Samples
Mean
95%
Max
L*
a*
b*
81
7.98
18.10
32.40
65.74
3.19
41.31
256
7.59
16.72
31.15
53.58
68.69
66.94
625
7.11
15.99
32.02
52.15
67.95
67.91
1296
7.07
16.03
33.33
51.77
68.32
69.21
2401
7.12
16.15
33.06
51.92
68.22
68.94
4096
6.96
15.68
38.30
50.91
69.73
74.10
6561
6.99
16.10
36.93
50.83
69.32
72.76
10000
7.06
16.22
36.71
51.19
69.36
72.54
Search WWH ::
Custom Search