Graphics Reference
In-Depth Information
ln( E i Δ t j )
ln( E i Δ t j )
4
4
2
2
0
0
-2
-2
-4
-4
pixel value Z ij
pixel value Z i j
-6
-6
0
50
100
150
200
250
0
50
100
150
200
250
(a) Original pixel values
(b) Recovered exposure values
Figure 6.2
Images of increasing exposure are used to reconstruct the response curve. (a) Plots of pixel
values from the original images. The different marks correspond to different pixels; values
across five images are plotted. (b) The corresponding recovered exposure values computed
from the reconstructed response curve. (After [Debevec and Malik 97].)
ln( E i Δ t j )
sample points
reconstructed
curve g
pixel value Z ij
Z min
Z mid
Z max
Figure 6.3
Response curve recovery: the response curve is fit to pixel values in the linear range,
weighted by how close they are to the middle of the range. (After [Debevec and Malik 97].)
6.1.4 HDRI and the Radiance Map
Given the recovered inverse response function g , the next step is to construct
the radiance map, which amounts to computing the values of E for each pixel.
The values of E i for the pixel samples are recovered in the process of minimiz-
 
Search WWH ::




Custom Search