Color in Image and Video
1. Suppose we use a new set of color-matching functions xnew(λ), ynew(λ), znew(λ) with values
|
λ(nm) |
xnew(λ) |
ynew(λ) |
znew(λ) |
|
450 |
0.2 |
0.1 |
0.5 |
|
500 |
0.1 |
0.4 |
0.3 |
|
600 |
0.1 |
0.4 |
0.2 |
|
700 |
0.6 |
0.1 |
0.0 |
In this system, what are the chromaticity values (x,y) of equi-energy white light E(λ) where E(λ) = 1 for all wavelengths λ? Explain.
2. Suppose image file values are in 0~255 in each color channel. If we define R=R/255 for the red channel, we wish to carry out gamma correction by passing a new value R’ to the display device, with R’≈R1/2.0.
It is common to carry out this operation using integer math. Suppose we approximate the calculation as creating new integer values in 0~255 via
(int) (255·(R1/2.0))
(a) Comment (very roughly) on the effect of this operation on the number of actually available levels for display. Hint: coding this up in any language will help you understand the mechanism at work better—and will allow you to simply count the output levels.
(b) Which end of the levels 0~255 is affected most by gamma correction—the low end(near 0) or the high end(near 255)? Why? How much at each end?