It is well known that the spectral sensitivities (SS) for color imaging devices should satisfy the Luther condition which requires SS to be linear combinations of the CIE color matching functions. In practice, it is difficult to construct an SS exactly to meet this condition. Some quality factors such as q-factor for single SS by Neugebauer and μ-factor for a set of SS by Vora et al. were introduced in order to describe the deviation of SS from their nearest color mixture curves. In this paper, a simple method is introduced to implement an evaluation platform for the above two quality factors. A series of hypothetical spectral sensitivities are constructed with cubic spline functions with shape and peak position of the SS parametrically varied. The evaluation platform is used to optimize these SS parameters to obtain a maximum quality factor. Furthermore, the improvement of quality factor by adding a fourth SS is discussed in the paper as well.
[1]
H. Joel Trussell,et al.
Mathematical method for designing a set of color scanning filters
,
1993,
Electronic Imaging.
[2]
Johji Tajima.
New Quality Measures for a Set of Color Sensors ∼ Weighted Quality Factor, Spectral Characteristic Restorability Index and Color Reproducibility Index ∼
,
1996,
Color Imaging Conference.
[3]
H. Neugebauer.
Quality Factor for Filters Whose Spectral Transmittances are Different from Color Mixture Curves, and Its Application to Color Photography*
,
1956
.
[4]
H. Joel Trussell,et al.
Measure of goodness of a set of color-scanning filters
,
1993
.
[5]
Gaurav Sharma,et al.
Figures of merit for color scanners
,
1997,
IEEE Trans. Image Process..
[6]
Michael J. Vrhel,et al.
Color correction using principal components
,
1992
.