Perceiving in depth Volume 2, Stereoscopic vision

642 Pages·2012·20.176 MB·English

by  Howard, Ian Porteous, Rogers, Brian J

PERCEIVING IN DEPTH
OXFORD PSYCHOLOGY SERIES
1. Th e Neuropsychology of Anxiety 18. Perceptual and Associative Learning 34. Looking Down on Human Intelligence
J. A. Gray G. Hall J. Deary
2. Elements of Episodic Memory 19. Implicit Learning and Tacit Knowledge 35. From Conditioning to Conscious
E. Tulving S. Reber Recollection
H. Eichenbaum and N. J. Cohen
3. Conditioning and Associative Learning 20. Neuromotor Mechanisms in Human
N. J. Mackintosh Communication 36. Understanding Figurative Language
D. Kimura S. Glucksberg
4. Visual Masking
B. G. Breitmeyer 21. Th e Frontal Lobes and Voluntary Action 3 7. Active Vision
R. Passingham M. Findlay and I. D. Gilchrist
5. Th e Musical Mind
J. A. Sloboda 22. Classifi cation and Cognition 38. Th e Science of False Memory
W. K. Estes C. J. Brainerd and V. F. Reyna
6. Elements of Psychophysical Th eory
J.-C. Falmagne 23. Vowel Perception and Production 39. Th e Case for Mental Imagery
B. S. Rosner and J. B. Pickering S. M. Kosslyn, W. L. Th ompson, and G. Ganis
7. Animal Intelligence
L. Weiskrantz 24. Visual Stress 40. Seeing Black and White
Wilkins Gilchrist
8. Response Times
R. D. Luce 25. Electrophysiology of Mind 41. Visual Masking, 2e
Edited by M. D. Rugg and M. G. H. Coles B. Breitmeyer and H. Öğmen
9. Mental Representations
Paivio 26. Attention and Memory 42. Motor Cognition
N. Cowan M. Jeannerod
10. Memory, Imprinting, and the Brain
G. Horn 27. Th e Visual Brain in Action 43. Th e Visual Brain in Action
D. Milner and M. A. Goodale D. Milner and M. A. Goodale
11. Working Memory
Baddeley 28. Perceptual Consequences of Cochlear 44. Th e Continuity of Mind
Damage M. Spivey
12. Blindsight
B. C. J. Moore
L. Weiskrantz 45. Working Memory, Th ought, and Action
29. Perceiving in Depth, Vols. 1, 2, and 3 Baddeley
13. Profi le Analysis
I. P. Howard with B. J. Rogers
D. M. Green 46. What Is Special about the Human Brain?
30. Th e Measurement of Sensation R. Passingham
14. Spatial Vision
D. Laming
R. L. DeValois and K. K. DeValois 47. Visual Refl ections
31. Conditioned Taste Aversion M. McCloskey
15. Th e Neural and Behavioural Organization
J. Bures, F. Bermúdez-Rattoni,
of Goal-Directed Movements 48. Principles of Visual Attention
and T. Yamamoto
M. Jeannerod C. Bundesen and T. Habekost
32. Th e Developing Visual Brain
16. Visual Pattern Analyzers 49. Major Issues in Cognitive Aging
J. Atkinson
N. V. S. Graham T. A. Salthouse
33. Th e Neuropsychology of Anxiety, 2e
17. Cognitive Foundations of Musical Pitch
J. A. Gray and N. McNaughton
C. L. Krumhansl
PERCEIVING IN DEPTH
VOLUME 2 STEREOSCOPIC VISION
Ian P. Howard Brian J. Rogers
CENTRE FOR VISION RESEARCH DEPARTMENT OF EXPERIMENTAL PSYCHOLOGY
YORK UNIVERSITY OXFORD UNIVERSITY
TORONTO
1
1

Oxford University Press, Inc., publishes works that further Oxford University’s objective of excellence
in research, scholarship, and education.
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur
Madrid Melbourne Mexico City Nairobi N ew Delhi Shanghai Taipei Toronto
With offi ces in
Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy
Japan Poland Portugal Singapore South Korea Switzerland Th ailand Turkey Ukraine Vietnam
Copyright © 2012 by Oxford University Press, Inc.
Published by Oxford University Press, Inc.
198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
____________________________________________

A copy of this book’s Cataloging-in-Publication Data is on fi le
with the Library of Congress.
ISBN: 978-0-19-976415-0
____________________________________________
9 8 7 6 5 4 3 2 1
Printed in the United States of America
on acid-free paper
CONTENTS OF VOLUME 2
11. Physiology of disparity detection 1 21. Depth contrast 433
12. Binocular fusion and rivalry 51 22. Stereopsis and perceptual organization 470
13. Binocular summation, masking, and transfer 107 23. Th e Pulfrich eff ect 515
14. Binocular correspondence and the horopter 148 24. Stereoscopic techniques and applications 538
15. Linking binocular images 182
16. Cyclopean vision 210 References 564
17. Stimulus tokens for stereopsis 249 Index of cited journals 621
18. Stereoscopic acuity 287 P ortrait index 624
19. Types of binocular disparity 363 Subject index 625
20. Binocular disparity and depth perception 385
v
A NOTE ON VIEWING THE STEREOGRAMS
Th e stereograms presented in this book can be fused with well worth acquiring, since it is oft en the only way to achieve
the aid of the prisms provided. Th e prisms should be held fusion with displays presented at vision conferences or
close to the eyes and about 12 cm above the page. Th e viewer when you have lost your stereoscope.
should be parallel to the plane of the page and correctly ori- A pair of images has one sign of disparity when fused by
ented within that plane. Incorrect orientation shows as an convergence and the opposite sign of disparity when fused
elevation of one image with respect to the other. Th e act of by divergence. A change in the sign of disparity reverses the
fusing a side-by-side pair of images by diverging the eyes is apparent depth relationships in the fused image. For stereo-
known as divergent fusion or uncrossed fusion. Th e act of grams in which the sign of disparity does not matter for the
fusing images by converging the eyes is c onvergent fusion illustration of a phenomenon, only one pair of images is
or crossed fusion. Th e prisms fuse the images by divergent provided. When the eff ect depends on the sign of disparity,
fusion. Stereograms may also be free-fused by diverging or two stereogram pairs are provided — one pair for readers
converging the eyes. In learning to free-fuse, it helps if the who prefer to converge the eyes, and the other for readers
eyes converge on a pencil point held at the correct distance who prefer to diverge. Note that the provided lenses fuse by
between the stereogram and the eyes. For divergence, one divergence only. Some stereograms in the book have triple
may place a piece of clear plastic over the stereogram and images in a row. Th ese create two side-by-side fused images
fi xate the refl ection of a point of light seen beyond the plane with opposite signs of disparity, plus fl anking monocular
of the stereogram. Th e correct distance can be found by images. Th erefore, four images are seen when the images are
observing how the images move as the pencil is moved in correctly fused. In some cases, it is instructive to compare
depth. the image formed by convergent fusion with that formed
Aft er some practice, readers will fi nd that they can con- by divergent fusion. In other cases, only one of the fused
verge or diverge the eyes without an aid. When stereograms images is of interest. In this case, the location of the
are free-fused, one sees each eye’s image on either side of the fused image of interest is indicated by a cross for those
fused image. Th e presence of three pictures confi rms that who fuse by convergence and by two parallel lines for those
correct vergence has been achieved. Free fusion is a skill who fuse by divergence.
vi
11
PHYSIOLOGY OF DISPARITY DETECTION
11.1 Introduction 1 11.5.1 Disparity detectors in V2 and V3 24
11.1.1 Basic terms 1 11.5.2 Disparity detectors in the dorsal stream 26
11.1.2 Discovery of disparity detectors 2 11.5.3 Disparity detectors in the ventral stream 28
11.2 Subcortical disparity-tuned cells 4 11.5.4 Parvo- and magnocellular disparity detectors 29
11.2.1 Disparity tuning in the pulvinar 4 11.6 Higher-order disparities 30
11.2.2 Disparity tuning in the nucleus of the optic tract 4 11.6.1 Detection of horizontal disparity gradients 31
11.2.3 Disparity tuning in the superior colliculus 4 11.6.2 Detection of vertical disparity gradients 31
11.3 Disparity detectors in cats 5 11.6.3 Spatial modulations of disparity 33
11.3.1 Disparity detectors in areas 17 and 18 5 11.6.4 Joint tuning to disparity and motion 33
11.3.2 Disparity detectors in higher visual areas of cats 6 11.6.5 Joint spatial and temporal disparities 34
11.4 Disparity detectors in primate V1 6 11.7 Evoked potentials and stereopsis 34
11.4.1 Disparity tuning functions 6 11.8 PET, fMRI, and stereopsis 37
11.4.2 Number and homogeneity of detectors 14 11.8.1 Stationary stimuli 37
11.4.3 Position- and phase-disparity detectors 16 11.8.2 Motion in depth 38
11.4.4 Detection of vertical disparity 18 11.9 Detection of midline disparity 39
11.4.5 Orientation and disparity tuning 20 11.9.1 Eff ects of midline section of the chiasm 39
11.4.6 Disparity tuning and eye position 21 11.9.2 Eff ects of callosectomy 40
11.4.7 Disparity in contrast-defi ned stimuli 22 11.10 Models of disparity processing 40
11.4.8 Dynamics of disparity detectors 23 11.10.1 Energy models 40
11.5 Disparity detection in higher visual centers 11.10.2 Neural network models 49
of primates 24
11.1 INTRODUCTION stimulus falls outside the fi eld of view of the other eye, or
(c) the stimulus is occluded to the other eye by a nearer
stimulus.
11.1.1 BASIC TERMS
A binocular stimulus is one seen at the same time by
Th e term “ stereoscopic vision ” means literally, “solid sight.” both eyes. Th e term “ dichoptic ” was originally used to
Strictly speaking, it refers to the visual perception of the describe the well-separated eyes of insects in contrast to
3-D structure of the world, when seen by one eye or by two. holoptic eyes, which have overlapping visual fi elds. Th e
However, the term is generally used to refer to 3-D depth term “dichotic” was coined by Stumpf ( 1916 ) to denote
perception arising from binocular disparities. the stimulation of each ear by a distinct sound. By analogy,
S everal terms referring to binocular vision are in the term “dichoptic” has come to mean stimulation of the
common use, but their meanings vary from author to two eyes by distinct distal stimuli (see Wade and Ono
author. Strictly speaking, all animals with two eyes have 2005 ).
binocular vision. Even animals with laterally placed eyes A dichoptic stimulus consists of distinct distal stimuli,
integrate the information from the two eyes to form a one presented to one eye and one to the other, which an
coherent representation of the fi eld of view. Also, the fi eld experimenter can control independently. Th ere are two
of view of almost all mammals has some region in which the basic procedures for gaining dichoptic control. Th e fi rst is
monocular fi elds overlap. However, the term “ binocular to present distinct stimuli to the two eyes in a stereoscope
vision ” is usually reserved for animals possessing a large area or by an equivalent procedure such as free fusion. Th e other
of binocular overlap within which diff erences between the procedure is to place diff erent fi lters or lenses in front of the
images are used to code depth. two eyes. Dichoptic stimuli usually diff er in some defi ned
A m onocular stimulus is a distal display seen by only way specifi ed by an experimenter. Th e diff erence may be
one eye because (a) the other eye is closed or absent, (b) the (a) a disparity of position, size, or orientation between parts
1
or the whole of the stimuli, or (b) a diff erence in luminance,
contrast, color, shape, or motion.
Th e term “ dioptic stimulus ” has been used to mean a
pair of identical stimuli in a stereoscope, in contrast to
dichoptic stimuli, which diff er in some respect (Gulick and
Lawson 1976 ). In the masking literature, a dichoptic stimu-
lus is one in which a mask and a test stimulus are shown
to diff erent eyes. In a dioptic stimulus they are shown to
both eyes, and in a monoptic stimulus they are shown to
one eye.
Th e term “ monoptic depth ” has been used to denote an
impression of depth created by a single eccentric stimulus in
one eye (Section 17.6.5).
11.1.2 DISCOVERY OF DISPARITY
DETECTORS
Before the 1960s many scientists, including Helmholtz,
believed that stereopsis does not involve conjunction of
inputs from the two eyes at an early stage (Section 2.10.5).
Ramón y Cajal ( 1911 ) proposed that inputs from corre-
sponding retinal regions converge on what he called “isody-
namic cells” and that this forms the basis of unifi ed
binocular vision. Th is idea was confi rmed when Hubel and
Wiesel ( 1959 , 1 962 ) reported that cells in the cat’s visual Figure 11.2. Colin Blakemore. Born in Stratford-upon-Avon, England in
1944. He obtained a B.A. in medical sciences from Cambridge
University in 1965 and a Ph.D. in physiological optics from Berkeley in
1968. He also holds a Sc.D. (Cantab) and D.Sc. (Oxon). He was
lecturer in physiology at Cambridge from 1972 to 1979, and then
Waynfl ete Professor of Physiology at Oxford University. Between 1990
and 2003 he was director of the McDonnell-Pew Center for Cognitive
Neuroscience in Oxford. He is now chief executive of the Medical
Research Council. He has been the recipient of the Robert Bing Prize
from the Swiss Academy of Medical Sciences, the Netter Prize from the
French Académie Nationale de Médecine, the Royal Society Michael
Faraday Award, the G.L. Brown Prize from the Physiological Society,
and the Charles F. Prentice Award from the American Academy of
Optometry.
cortex receive inputs from the two eyes and that the recep-
tive fi elds of these b inocular cells occupy corresponding
positions in the two eyes. But consider what would happen
if the monocular receptive fi elds of each binocular cell
occupied identical positions in the retinas and were identi-
cal in all other respects. Each binocular cell would respond
optimally to similar stimuli with zero disparity. Such cells
would not be diff erentially tuned to diff erent disparities
and would therefore be incapable of coding relative depth.
If images falling on corresponding locations diff ered, inputs
to binocular cells could sum or rival. For stereopsis, the
visual system needs binocular cells that are maximally
responsive to inputs from receptive fi elds that are displaced
Figure 11.1. Horace B. Barlow. Born in England in 1921. He graduated
by diff erent amounts from exact correspondence. Any such
from Trinity College, Cambridge. He was a research fellow at Trinity
College between 1950 and 1954 and a lecturer at King’s College, cell would respond optimally to a stimulus with disparity
Cambridge, between 1954 and 1964. Between 1964 and 1973 he was of a given magnitude and sign (crossed or uncrossed). A set
professor of physiological optics and physiology at the University of
of such cells with diff erent receptive-fi eld off sets in one
California at Berkeley. He then returned to the physiological laboratory
sign or the other could code diff erent disparities and hence
at Cambridge University as a Royal Society Research Professor. He
became a fellow of the Royal Society of London in 1969. relative depth.
2 • STEREOSCOPIC VISION
In 1967 Jack Pettigrew discovered binocular cells with
these properties in the cat’s visual cortex. Such cells are now
known as disparity detectors. Pettigrew was a student
of Peter Bishop at Sydney University, Australia (Section
2.10.5). He then joined Barlow and Blakemore in Berkeley,
California. Working together, they confi rmed the existence
of disparity detectors in the cat (Barlow et al. 1967 ) (Portrait
Figures 11.1 , 11.2 , and 11.3 ).
Similar fi ndings were reported about the same time
from Sydney by Pettigrew, Nikara, and Bishop ( 1968 )
(Portrait Figure 11.4) . In 1977, Gian Poggio and his
coworkers at Johns Hopkins University discovered dispar-
ity detectors in monkey V1.
Th e search for binocular cells tuned to diff erent dispari-
ties was beset with the problem of ensuring that the images
in the two eyes are in register. If the images are slightly out
of register, a cell tuned to zero disparity will appear to be
tuned to a disparity equal to the image misregistration.
Also, any movement of the eyes during the recording intro-
duces artifacts. Several procedures have been used to solve
this problem. In the anesthetized animal, eye movements Figure 11.4. Peter O. Bishop. Born in Tamworth, New South Wales,
are controlled by paralyzing the eye muscles and attaching Australia in 1917. He obtained the M.B. and B.S. in 1940 and the D.Sc.
the eyeball to a clamped ring. A rotating mirror or a prism in 1967 from the University of Sydney. Aft er serving as a surgeon
during the war he studied at University College London. He held
of variable power controls the eff ective direction of gaze. In
academic appointments at the University of Sydney from 1950 to 1967
the reference-cell procedure, diff erent electrodes record when he became professor of physiology at the Australian National
University in Canberra. He retired in 1983. He is a fellow of the
Australian Academy of Sciences, fellow of the Royal Society of London,
offi cer of the Order of Australia, and joint winner of the Australia Prize
in 1993.
responses of a test cell and a reference binocular cell, each
with receptive fi elds in the central retinas. Changes in the
response of the reference cell indicate when eye movements
have occurred (Hubel and Wiesel 1970 ). In a related proce-
dure, eye drift is monitored by the response of a reference
cell to monocular stimulation (Maske et al. 1986a ). Image
stability can also be indicated by responses of LGN cells of
foveal origin, one from each eye (LeVay and Voigt 1988 ).
Th ese procedures indicate when eye drift has occurred,
but they do not specify when test stimuli have zero dispar-
ity, since the reference cell may not be tuned to zero dispar-
ity. One solution to this problem is to use the mean response
of several reference cells to defi ne zero disparity (Nikara
et al. 1968 ). Another procedure is to use an ophthalmo-
scope to project images of retinal blood vessels onto the
screen on which the stimuli are presented (Bishop et al.
1962 ; Pettigrew et al. 1 968) . Th e problem is simplifi ed
when testing is done on alert monkeys trained to converge
Figure 11.3. John D. Pettigrew. Born in Wagga Wagga, Australia, in 1943. their eyes on defi ned targets.
He obtained his B.Sc. in physiology from the University of Sydney in
I n identifying a disparity detector one must ensure that
1966 and an M.B. from Sydney University Medical School in 1969.
He has worked at Berkeley with H. Barlow and C. Blakemore, at the changes in responses are not due to incidental changes in
California Institute of Technology, Queens University in Canada, and stimulation. For example, the stimulus in one eye may move
the Zoologisches Insititüt in Munich. Since 1988 he has been director,
outside the receptive fi eld of the binocular cell when the
of the Vision, Touch, and Hearing Research Centre at the University of
experimenter changes the disparity of the stimuli. Eff ects
Queensland, Australia. He is a fellow of the Royal Society of London
and of the Australian Academy of Science. of monocular position can be separated from eff ects of
PHYSIOLOGY OF DISPARITY DETECTION • 3