1. Digital Visual Effects, Spring 2006 Yung-Yu Chuang 2006/3/8
Tone mapping
Digital Visual Effects, Spring 2006
Yung-Yu Chuang
2006/3/8
Basics for cameras
Some things you might want to know when you shoot pictures
First time, not well-organized yet
Scribe
Assignments
Slides available before Thursday
with slides by Fredo Durand, and Alexei Efros

2. Tone mapping 10-6 106 10-6 106 Pixel value 0 to 255
How can we display it?
Linear scaling?, thresholding?
10-6
106
dynamic range
Real world
radiance
10-6
106
Display
intensity
Pixel value 0 to 255
CRT has 300:1 dynamic range

3. Global operator (Reinhart et al)

4. Global operator results

5. Eye is not a photometer!
"Every light is a shade, compared to the higher lights, till you come to the sun; and every shade is a light, compared to the deeper shades, till you come to the night."
— John Ruskin, 1879

6. Compressing dynamic range

7. Fast Bilateral Filtering for the Display of High-Dynamic-Range Images
Frédo Durand & Julie Dorsey
Laboratory for Computer Science
Massachusetts Institute of Technology

8. A typical photo
Sun is overexposed
Foreground is underexposed

9. Gamma compression
X -> Xg
Colors are washed-out
Input
Gamma

10. Gamma compression on intensity
Colors are OK, but details (intensity high-frequency) are blurred
Intensity
Gamma on intensity
Color

11. Chiu et al. 1993 Reduce contrast of low-frequencies
Keep high frequencies
Low-freq.
Reduce low frequency
High-freq.
Color

12. The halo nightmare For strong edges
Because they contain high frequency
Low-freq.
Reduce low frequency
High-freq.
Color

13. Durand and Dorsey Do not blur across edges Non-linear filtering
Large-scale
Output
Detail
Color

14. Edge-preserving filtering
Blur, but not across edges
Anisotropic diffusion [Perona & Malik 90]
Blurring as heat flow
LCIS [Tumblin & Turk]
Bilateral filtering [Tomasi & Manduci, 98]
Input
Gaussian blur
Edge-preserving

15. Start with Gaussian filtering
Here, input is a step function + noise
output
input

16. Start with Gaussian filtering
Spatial Gaussian f
output
input

17. Start with Gaussian filtering
Output is blurred
output
input

18. Gaussian filter as weighted average
Weight of x depends on distance to x
output
input

19. The problem of edges Here, “pollutes” our estimate J(x)
It is too different
output
input

20. Principle of Bilateral filtering
[Tomasi and Manduchi 1998]
Penalty g on the intensity difference
output
input

21. Bilateral filtering Spatial Gaussian f output input
[Tomasi and Manduchi 1998]
Spatial Gaussian f
output
input

22. Bilateral filtering Spatial Gaussian f
[Tomasi and Manduchi 1998]
Spatial Gaussian f
Gaussian g on the intensity difference
output
input

23. Normalization factor
[Tomasi and Manduchi 1998]
k(x)=
output
input

24. Bilateral filtering is non-linear
[Tomasi and Manduchi 1998]
The weights are different for each output pixel
output
input

25. Contrast reduction
Input HDR image
Contrast too high!

26. Contrast reduction
Input HDR image
Intensity
Color

27. Contrast reduction Large scale Intensity Fast Bilateral Filter
Input HDR image
Large scale
Intensity
Fast Bilateral Filter
Color

28. Contrast reduction Large scale Fast Bilateral Filter Detail
Input HDR image
Large scale
Intensity
Fast Bilateral Filter
Detail
Color

29. Contrast reduction Scale in log domain Large scale Large scale
Input HDR image
Scale in log domain
Large scale
Large scale
Intensity
Reduce contrast
Fast Bilateral Filter
Detail
Color

30. Contrast reduction Large scale Large scale Reduce contrast
Input HDR image
Large scale
Large scale
Intensity
Reduce contrast
Fast Bilateral Filter
Detail
Detail
Preserve!
Color

31. Contrast reduction Output Large scale Large scale Reduce contrast
Input HDR image
Output
Large scale
Large scale
Intensity
Reduce contrast
Fast Bilateral Filter
Detail
Detail
Preserve!
Color
Color