71
Computational Theories & Low- level Pixels To Percepts A. Efros, CMU, Spring 2009

Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

  • View
    225

  • Download
    2

Embed Size (px)

Citation preview

Page 1: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Computational Theories & Low-level

Pixels To PerceptsA. Efros, CMU, Spring 2009

Page 2: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

Image- BasedProcessing

Surface- BasedProcessing

Object-Based

Processing

Category- BasedProcessing

Light

Vision

Audition

STM

LTM

Motor

Sound

LightMove-ment

Odor (etc.)

Ceramiccup on a table

David Marr, 1982

Page 3: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

The Retinal Image

An Image (blowup) Receptor Output

Page 4: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

Image-basedRepresentation

Primal Sketch(Marr)

An Image

(Line Drawing)

RetinalImage

Image-based

processes

EdgesLinesBlobsetc.

Page 5: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

We likely throw away a lot

Page 6: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

line drawings are universal

Page 7: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

Surface-basedRepresentation

Primal Sketch 2.5-D Sketch

Image-basedRepresentation

Surface-based

processes

StereoShadingMotion

etc.

Page 8: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Single Surface(Koenderink’s trick)

Page 9: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

Surface-basedRepresentation

Primal Sketch 2.5-D Sketch

Image-basedRepresentation

Surface-based

processes

StereoShadingMotion

etc.

Page 10: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Figure/Ground Organization

A contour belongs to one of the two (but not both) abutting regions.

Figure(face)

Ground(shapeless)

Figure(Goblet)Ground

(Shapeless)

Important for the perception of shape

Page 11: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Properties of figures vs. grounds

15.18

Figure GroundThing-like Not thing-likeCloser FartherShaped Extends behind

Figure-Ground OrganizationFigure-Ground Organization

Page 12: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Surroundedness

15.19Figure-Ground OrganizationFigure-Ground Organization

Surrounded region --> FigureSurrounding region --> Ground

Page 13: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Size

15.20Figure-Ground OrganizationFigure-Ground Organization

Smaller region --> FigureLarger region --> Ground

Page 14: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Orientation

15.21Figure-Ground OrganizationFigure-Ground Organization

Horizontal/vertical region --> FigureOblique region --> Ground

Page 15: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Contrast

15.22Figure-Ground OrganizationFigure-Ground Organization

Higher contrast region --> FigureLower contrast region --> Ground

Page 16: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Symmetry

15.23Figure-Ground OrganizationFigure-Ground Organization

Symmetrical region --> FigureAsymmetrical region --> Ground

Page 17: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Convexity

15.24Figure-Ground OrganizationFigure-Ground Organization

More convex region --> FigureLess convex region --> Ground

Page 18: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Parallelism

15.25Figure-Ground OrganizationFigure-Ground Organization

More parallel region --> FigureLess parallel region --> Ground

Page 19: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Lower region

15.26Figure-Ground OrganizationFigure-Ground Organization

Lower region --> FigureUpper region --> Ground

Page 20: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Meaningfulness

15.27Figure-Ground OrganizationFigure-Ground Organization

More meaningful region --> FigureLess meaningful region --> Ground

Page 21: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Relation to Depth Factors

15.28Figure-Ground OrganizationFigure-Ground Organization

Figure-ground organization as edge assignment:To which side does the edge belong?

Depth cues can also be figure-ground factorsand

Figure-ground factors can be depth cues.

To the closer side. This fact connects figure-groundorganization with depth perception.

Page 22: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Occlusion

15.29Figure-Ground OrganizationFigure-Ground Organization

Occluding region --> FigureOccluded region --> Ground

Page 23: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Cast Shadows

15.30Figure-Ground OrganizationFigure-Ground Organization

Shadowing region --> FigureShadowed region --> Ground

Page 24: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

Principles of figure-ground organization:

Shading

15.32Figure-Ground OrganizationFigure-Ground Organization

Shaded region --> FigureNonshaded region --> Ground

Page 25: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Line Labeling

> : contour direction+ : convex edge - : concave edge

possible junctions(constraints)

ConstraintPropagation

[Clowes 1971, Huffman 1971; Waltz 1972; Malik 1986]

Page 26: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

26

Page 27: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Line Labeling

Page 28: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

Object-basedRepresentation

Object-based

processes

GroupingParsing

Completionetc.

Surface-basedRepresentation

2.5-D Sketch Volumetric Sketch

Page 29: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Geons(Biederman '87)

Page 30: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Four Stages of Visual PerceptionFour Stages of Visual Perception

© Stephen E. Palmer, 2002

Category-basedRepresentation

Category-based

processes

Pattern-Recognition

Spatial-description

Object-basedRepresentation

Volumetric Sketch Basic-level Category

Category: cup

Color: light-gray

Size: 6”

Location: table

Page 31: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

We likely throw away a lot

Page 32: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

line drawings are universal

Page 33: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

However, things are not so simple…

● Problems with feed-forward model of processing…

Page 34: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Junctions in Real Images

Page 35: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Are Junctions local evidence?

J McDermott, 2004

Page 36: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.38

Is grouping an early or late process?

Early vs. Late GroupingEarly vs. Late Grouping

Image- BasedProcessing

Surface- BasedProcessing

Object-Based

Processing

Category- BasedProcessing

Light ? ? ? ?

Page 37: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.39

Before or after stereoscopic depth?

(Rock & Brosgole, 1964)

Early vs. Late GroupingEarly vs. Late Grouping

Page 38: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.40

Before or after lightness constancy?

(Rock, Nijhawan, Palmer & Tudor, 1992)

ReflectanceMatched

LuminanceMatched

TranslucentPlastic Strip

Early vs. Late GroupingEarly vs. Late Grouping

ReflectanceMatched

Luminance-Ratio Matched

OpaquePaper Strip

Opaquepaper strip

Page 39: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.41

Before or after visual completion?

(Palmer, Neff & Beck, 1996)

Early vs. Late GroupingEarly vs. Late Grouping

Page 40: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.42

Before or after illusory contours?

(Palmer & Nelson, 2000)

?

Early vs. Late GroupingEarly vs. Late Grouping

Page 41: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.43

Conclusion: Grouping can occur “late”

Question: Can grouping also occur “early”

(Palmer & Brooks, in preparation)

Early vs. Late GroupingEarly vs. Late Grouping

Page 42: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.44

Grouping affects shape constancy

(Palmer & Brooks, in preparation)

Ambiguous

Flat oval

Circle in depth

Early vs. Late GroupingEarly vs. Late Grouping

Page 43: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.45

Proximity effects

Biased toward oval

Biased toward circle

Early vs. Late GroupingEarly vs. Late Grouping

Page 44: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.46

Color similarity effects

Biased toward oval Biased toward circle

Early vs. Late GroupingEarly vs. Late Grouping

Page 45: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.47

Common fate effects

Biased toward oval Biased toward circle

Early vs. Late GroupingEarly vs. Late Grouping

Page 46: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

© Stephen E. Palmer, 2002

14.48

Conclusion: Grouping occurs both “early”

and “late” -- possibly everywhere!

Image- BasedProcessing

Surface- BasedProcessing

Object-Based

Processing

Category- BasedProcessing

Light

Grouping Grouping Grouping Grouping

Early vs. Late GroupingEarly vs. Late Grouping

Page 47: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

two-tone images

Page 48: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009
Page 49: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009
Page 50: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

hair (not shadow!)

inferred external contours

“attached shadow” contour

“cast shadow” contour

Page 51: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009
Page 52: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Finding 3D structure in two-tone images requires distinguishing cast shadows, attached shadows, and areas of low reflectivity

The images do not contain this information a priori (at low level)

Cavanagh's argument

Page 53: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

A Classical View of Vision

Grouping /Segmentation

Figure/GroundOrganization

Object and Scene Recognition

pixels, features, edges, etc.Low-level

Mid-level

High-level

Page 54: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

A Contemporary View of Vision

Figure/GroundOrganization

Grouping /Segmentation

Object and Scene Recognition

pixels, features, edges, etc.Low-level

Mid-level

High-level

But where we draw this line?

Page 55: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Question #1:What (if anything) should be done at the “Low-Level”?

N.B. I have already told you everything that is known. From now on, there

aren’t any answers.. Only questions…

Page 56: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Who cares? Why not just use pixels?

Pixel differences vs. Perceptual differences

Page 57: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

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

Page 58: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Cornsweet Illusion

Page 59: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Campbell-Robson contrast sensitivity curveCampbell-Robson contrast sensitivity curve

Sine wave

Page 60: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Metamers

Page 61: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Question #1:What (if anything) should be done at the “Low-Level”?

i.e. What input stimulus should we be invariant to?

Page 62: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Invariant to:

• Brightness / Color changes?

small brightness / color changeslow-frequency changes

But one can be too invariant

Page 63: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Invariant to:

• Edge contrast / reversal?

I shouldn’t care what background I am on!

but be careful of exaggerating noise

Page 64: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Representation choices

Raw Pixels

Gradients:

Gradient Magnitude:

Thresholded gradients (edge + sign):

Thresholded gradient mag. (edges):

Page 65: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Spatial invariance

• Rotation, Translation, Scale• Yes, but not too much…

• In brain: complex cells – partial invariance

• In Comp. Vision: histogram-binning methods (SIFT, GIST, Shape Context, etc) or, equivalently, blurring (e.g. Geometric Blur) -- will discuss later

Page 66: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Many lives of a boundary

Page 67: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Often, context-dependent…

input canny human

Maybe low-level is never enough?

Page 68: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

1/f amplitude spectra for natural images

(Field 1987)

There are statistical regularities in the natural world, and image statistics reflect that. (Burton & Moorehead 1987; Field 1987; Tolhurst et al. 1992)

Page 69: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Why 1/f?

Scale invariance

Edges have 1/f structure

Object distribution in real world (Ruderman 1997; Lee & Mumford 1999)

(Image source: smokiesguidebook.comSlide content: Simoncelli & Olshausen 2001)

Page 70: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

A closer look at amplitude spectra

(Torralba & Oliva 2003)

Page 71: Computational Theories & Low-level Pixels To Percepts A. Efros, CMU, Spring 2009

Do natural image statistics matter?Sensory coding might exploit statistical regularities of our world according to various criteria:

Representational efficiency Decorrelate input responses, make them independent, sparse,

information theoretic metrics etc.

Metabolic efficiencySpike efficiency, minimal wiring.

Learning efficiencySparseness, invariance, over completeness etc.

Lots and lots of work; see reviews Graham & Field (2007), Simoncelli & Olshausen (2001)Lots and lots of work; see reviews Graham & Field (2007), Simoncelli & Olshausen (2001)