Depth perception
1
Images removed due to copyright restrictions.
Please refer to lecture video
2
Cues used for coding depth in the brain
Oculomotor cues
accommodation
vergence
Visual cues
Binocular
stereopsis
Monocular
motion parallax
shading
interposition
size
perspective
3
Stereopsis, basic facts and demos
4
Two simple methods for creating stereo images
1
stereo camera
2
swiveled regular camera
5
Image is in public domain.
Demos, page 1
6
Images removed due to copyright restrictions.
Please see lecture video or Figure 1 of Schiller, Peter H., Geoffrey L. Kendall, et al. "Depth
Perception, Binocular Integration and Hand-Eye Coordination in Intact and Stereo Impaired
Human Subjects." Journal of Clinical & Experimental Ophthalmology 3, (210).
Demos, page 1 & 2
7
Autostereogram from Magic Eye
Images removed due to copyright restrictions.
Please see lecture video or the autostereogram from The Magic Eye, Volume I:
A New Way of Looking at the World. Andrews McMeel Publishing, 1993.
Autostereograms: see Howard and Rogers, Depth Perception, Vol 2, pp. 549-556.
Demos, page 3
8
Retinal disparity utilized for stereopsis
Target hits corresponding retinal points
Target hits non-corresponding retinal points
target
target
fix
fix
Vieth-Muller circle
y
x
x=y
y
x
x=y
9
Stereopsis, neuronal responses
10
V1 cell response to drifting bars at varied stereo depths
0.58
+6.0
+4.0
0.21
+2.0
+0.5
_ 0
38 +
0
-0.5
Depth (cm)
Disparity (degree)
+1.5
+1.0
-1.0
-1.5
-0.23
-2.0
-4.0
-6.0
-1.13
-8.0
O
-10
3
O
0
1
2
sec
O
0.14
Image by MIT OpenCourseWare.
11
V1 cell response to drifting bars at varied stereo depths
+10.0
0.88
+8.0
+6.0
+4.0
+3.0
+2.0
+1.0
+1.0
38 +_ 0
0
Depth (cm)
Disparity (degree)
0.31
-0.5
-1.0
-0.23
-2.0
-4.0
-6.0
-1.13
-8.0
0
4
O
0
1
R
2
sec
3
o
0.18
38
L
Key down
Image by MIT OpenCourseWare.
12
Basic cell types for stereo
Tuned excitatory near
Tuned excitatory zero
100
150
100
125
75
100
Neural responses (impulses per sec)
75
75
50
50
50
25
25
0
L
R
.05
00
05
10
.10
.05
00
10
05
100
75
Tuned inhibitory
Near
25
0
0
.10
100
Tuned excitatory far
.05 00
05
00
05
Far
75
75
50
50
50
25
25
25
0
.10
.05
00
05
10
0
0
.10
.05
00
05
10
Horizontal display (deg)
.10
.05
10
Image by MIT OpenCourseWare.
13
Basic cell types for stereo
Tuned excitatory
Near
Neural activity
]
Far
]
Tuned inhibitory
Near
Fixation point
Far
Depth
Image by MIT OpenCourseWare.
14
The effects of V4 and MT lesions
on stereoscopic depth perception
15
Images removed due to copyright restrictions.
Please see lecture video or Figure 1 of Schiller, Peter H., Geoffrey L. Kendall, et al. "Depth
Perception, Binocular Integration and Hand-Eye Coordination in Intact and Stereo Impaired
Human Subjects. "Journal of Clinical & Experimental Ophthalmology 3, (210).
16
Stereopsis
Graphs removed due to copyright restrictions.
Please see lecture video or Schiller, Peter H. "The Effects of V4 and Middle
Temporal (MT) Area Lesions on Visual Performance in the Rhesus Monkey."
Visual Neuroscience 10, no. 4 (1993): 717-46.
17
Motion parallax
18
MOTION PARALLAX, the eye is stationary
A
B
C
Velocity of motion:
A fastest, C slowest, creating velocity gradient
With rigidity constraint highest relative velocity is judged to be closest
19
MOTION PARALLAX, the eye tracks
1
a
2
b
eye movement
a
object motion
1
The eye tracks the circle, which
therefore remains stationary on the fovea
Objects nearer than the one tracked move
at greater velocities on the retinal surface
than objects further; the further objects
actually move in the opposite direction
on the retina.
2
b
20
21
Motion parallax
To derive depth information from motion parallax neurons are
needed that provide information about velocity and direction of
motion and perhaps also about differential motion.
The majority of cells in V1 are direction and velocity selective.
Some appear also to be selective for differential motion.
Cells that process motion parallax have also been found in MT.
22
Brain activation by stereopsis and
motion parallax in normal and
stereoblind subjects with fMRI
23
Images are
projected here
24
25
A: Stereo
27
B: Parallax
28
C: Stereo + Parallax
29
Shading
30
31
32
The power of shading for depth perception
33
Examples of the 12 different shadings used
34
The next five displays have identical,
repeating elements which are shaded
differently in each set yielding a variety
of stable and conflicting percepts
35
36
37
38
39
40
Stereo and shading in harmony and in conflict
Demos, page 4-7
41
f
f
f
42
43
44
Putting stereo, parallax and shading together
45
Image removed due to copyright restrictions.
Please see lecture video or see Display 24 of Schiller, Peter H., and Christina E.
Carvey. "Demonstrations of Spatiotemporal Integration and What they Tell us
About the Visual System." Perception 35, no. 11 (2006): 1521.
46
Separate and combined disparity, parallax and shading depth cues
Graphs removed due to copyright restrictions.
Please see lecture video or Figure 3a,b of Schiller, Peter H., Warren M. Slocum,
et al. "The Integration of Disparity, Shading and Motion Parallax Cues for Depth
Perception in Humans and Monkeys." Brain Research 1377 (2011): 67-77.
47
Perspective
48
Image removed due to copyright restrictions.
Please see lecture video or see William O'Brian, "Because it's
here, that's why," New Yorker, September 21, 1963, 42.
49
50
51
52
53
Perspective illusion
54
55
The Savage family
Painting is in public domain.
Edward Savage
56
Hand-eye coordination and
Binocular integration tests
57
Needle test
58
Movie of needle test
needle_binoc
needle_monoc
59
Touch-panel test
60
Examples of test of binocular integration
61
Binocular integration test, figures
A
B
62
Binocular integration test, words
A
S U D
B
ST UR D Y
T
R
Y
ST UR D Y
63
Summary, depth:
1. Numerous mechanisms for analyzing depth have been identified that
include vergence and accommodation, stereopsis, parallax, shading, and
perspective.
2. Several cortical structures process stereopsis utilizing disparity
information; the number of disparities represented is limited as
in the case of color coding.
3. Utilizing motion parallax for depth processing necessitates neurons
specific for direction, velocity and differential velocity; several areas,
including V1 and MT process motion parallax.
4. Area MT contributes to the analysis of motion, motion parallax, depth,
and flicker; however, these analyses are also carried out by several
other structures.
5. Litte is known at present about the manner in which information about
shading and perspective is analyzed by the brain.
64
MIT OpenCourseWare
http://ocw.mit.edu
9.04 Sensory Systems
Fall 2013
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.