1) Traditional color constancy algorithms can only handle the images which intensities are linear with the scene reflected illuminations.

2) Color constancy refers to the invariability of color perception of the same surface under different illuminations.

3) Furthermore, the degree of color constancy fluctuated with surface variation in the surface matching test,(http:///color constancy.html) which showed that color constancy was also dependent on the characteristics of the surface.

4) The degree of color constancy in the same background was higher than that in a different background.

5) It is essential for realizing color constancy to acquire color descriptors that are not affected by illumination and geometry conditions.

6) This study verifies and discusses color constancy with experimentation in terms of human vision and design.

7) Therefore they are not practical. In this dissertation, we provide a new color constancy algorithm which can handle uncalibrated images and is free of camera and shot parameters.

8) In this paper, we analyze the variation of saturation, also we compare the previous methods of color constancy and point out that they can not get the constant saturation information.

9) This description of the surface spectral reflectance has the property of color constancy.

10) To get a color descriptor that is robust to illumination and has the color constancy function is very important for the entire computer vision system.

11) The disinhibitory IF of subcortical neurons contributes to transmission of low spatial frequency information of visual images and tO the maintenance of color constancy.