1. L
Differentiate
Threshold
Integrate R

2. The lightness scaling problem:
Q. How should we assign an absolute lightness to a surface?
A. Anchoring – brightest region in field of view is declared to be ‘white’
Demo

3. A major open challenge for lightness perception models:
Moving beyond a flat world; distinguishing between abrupt orientation
changes and reflectance edges.
Sinha, 1992; Sinha and Adelson, 1993

4. Processing Framework Proposed by Marr
Recognition
3D structure; motion characteristics; surface properties
Shape
From
stereo
Motion
flow
Shape
From
motion
Color
estimation
Shape
From
contour
Shape
From
shading
Shape
From
texture
Edge extraction
Image

5. Shape from contour
I’m sorry, Mr. Mitchell is only painted on the wall.

6. The perceptual task 3D shape recovery from a single 2D image

9. A constraint based system
Symmetry

10. An implemented constraint-based system
Attneave & Frost, 1969
Mackworth, 1973
Perkins & Cooper, 1980
Kanade, 1980
Barrow & Tenenbaum, 1981
Marill, 1989
Fischler & Leclerc, 1992
A 3D shape
recovery
system with
3 general
constraints
but no
object-specific
learning
Experimentation
Sinha and Adelson
Occam’s razor and
complexity

11. The system in action...
2D Inputs
Recovered
3D shapes

12. The system in action...
Video

13. Where the system works...

14. And where it runs into trouble...

15. “Here be dragons” Natural objects Arbitrary polyhedra
Semi-regular polyhedra
Polyhedra with
rectangular facets
Circular discs

16. High-level learning based hypothesis

17. 3D structure; motion characteristics; surface properties
Recognition
3D structure; motion characteristics; surface properties
Shape
From
stereo
Motion
flow
Shape
From
motion
Color
estimation
Shape
From
contour
Shape
From
shading
Shape
From
texture
Edge extraction
Image

18. Experimental studies of object-specific
influences on early perception and a
simple model…
C:/BCS_core/vis_learning2.ppt