Transcript Ch 3 Vision - Texas A&M University

Sensation & Perception
• Ch. 3: Vision
© Takashi Yamauchi (Dept. of Psychology, Texas A&M University)
• Main topics
– convergence
– Inhibition, lateral inhibition and lightness perception
– Interactions between neurons
– Feature detectors
ch 3
1
Question 1
• What do these devices have in common?
ch 3
2
These devices make use of
electromagnetic waves
Capture electromagnetic waves and transform
them into electricity.
ch 3
3
What does the eye do?
 Transducing light energy into
electrical energy
ch 3
4
Transduction
•
•
•
•
•
 Light enters the eye
 A photon hits a receptor
 changes the shape of pigment molecules
 triggers massive chemical reactions
 generate electrical signals
ch 3
5
ch 3
6
• How big are they?
– Power of Ten (10 min)
– http://www.youtube.com/watch?feature=player
_embedded&v=0fKBhvDjuy0#!
ch 3
7
• Solar cells
(photovoltaics)
produce electricity in a
similar way as our
eyes do.
ch 3
8
Rods and cones
•
•
•
•
Morphology
Distribution on the retina
Dark adaptation
Spectral sensitivity
ch 3
9
Photo receptors: Rods and cones
• Rods have bigger outer segments than cones.
• Why?
ch 3
10
Outer segments capture photons
• Bigger outer segments can capture more light.
• Rods have bigger outer segments.
– Rods allow us to see in the dark.
– Cones are mainly for day vision.
– Cones are for color perception.
ch 3
11
How can we see objects?
• How can we see a book?
• How can we see a desk?
• Why don’t we see light?
ch 3
12
Reflection of light
• What we see is a reflection of light.
• Different objects reflect different
wavelengths,
–  different objects show different colors
ch 3
13
• Photo receptors in the eye are geared to
capture different wavelengths
ch 3
14
Lens: focuses light rays.
Iris: control the size of the pupil  regulating the
amount of light reaching the retina
Retina: a layer of receptor cells
Receptor cells  rods and cones
ch 3
15
Retina:
ch 3
16
• Photo receptors are facing away from the light source.
• The optic nerve carries neural information to this spot.
• What happens?
– No receptors, no vision
ch 3  blind spot
17
Some messages: how to improve your
vision
• Massage your eye muscles
• Eat carrots
• Massage the back of your head.
ch 3
18
Rods and cones
•
•
•
•
Morphology
Distribution on the retina
Dark adaptation
Spectral sensitivity
ch 3
19
The distribution of cones and rods
on the retina
• Cones are concentrated mainly
on the fovea.
• There are no rods on the fovea.
• We move eyes to capture images
on the fovea.
ch 3
20
Demonstration
• Blind spot
ch 3
21
Rods and cones are different
• In their dark adaptation rates
ch 3
22
Dark adaptation rates of rods and
cones
• When you enter a dark room from outside,
you can’t see well at first. But gradually,
your eyes are adjusted to the dark, and see
better.
ch 3
23
Review
ch 3
24
Visual perception
• What is the difference
between (a) & (b)?
(a)
• What is going on in your
head when you see (a)
versus when you see (b)?
(b)
ch 3
25
ch 3
26
ch 3
27
How about this?
ch 3
28
ch 3
29
ch 3
30
ch 3
31
ch 3
32
ch 3
33
ch 3
34
What’s going on?
• When you see the square, what’s going on?
• How do you find out?
ch 3
35
• In terms of the activity of neurons,
what is the difference between
A and B ?
Any guess?
B.
A.
ch 3
36
Measuring the electrical
activity of a neuron
directly by inserting a
thin needle into animal
brains.
ch 3
37
The frequency of
action potential
0
0
The number of
action potential
5 units
emitted by a
neuron is
correlated with
t
Time
the intensity of
the stimulus.
10 units
t
Time
20 units
ch 2
0
38
t
Time
Physical quantities
Perceived
quantities
5 units
0
t
Time
20 units
ch 2
0
t
Time
39
Neural Processing by
Convergence
• Why are rods more sensitive to light than
cones?
ch 3
40
Activities of neurons can be
schematically shown as
a1 a2 a3 a4
The firing rate of neuron
B is determined by the
activation sent by neurons
a1-a4.
B
ch 3
41
Ganglion cell
• Ganglion cell
ch 3
42
• Convergence:
• The ratio of connections with two groups of
neurons.
• Rods vs. Ganglion cells
– 120:1
• Cones vs. Ganglion cells
– 6:1
ch 3
43
ch 3
44
ch 3
45
Why does this matter?
• How is this related to the higher sensitivity
of rods?
ch 3
46
The cones result in better detail
vision than the rods
• Visual acuity
– How far apart are two dots?
ch 3
47
The frequency of
action potential
0
t
0
t
The number of
action potential
emitted by a
neuron is
correlated with
Time
the intensity of
the stimulus.
Time
ch 3
0
48
t
Time
ch 3
49
ch 3
50
ch 3
51
Fig. 2.11, p.53
ch 3
52
ch 3
53
ch 3
54
• Demonstration:
• On a scratch paper, draw two vertical
lines of about 2 inches (1/2 inch apart).
• Close your left eye, and focus your
right eye on your index figure, and
move the figure.
• At some point, you can’t distinguish
the two vertical lines.
ch 3
55
The distribution of cones and
rods on the retina
• Cones are concentrated mainly on the fovea.
• There are no rods on the fovea.
• We move eyes to capturech 3images on the fovea.
56
Visual Cortex
ch 3
57
ch 3
58
ch 3
59
The cones result in better detail
vision than the rods
• Visual acuity
– How far apart are two dots?
ch 3
60
Neurons
• How do you detect there are two separate
dots (lights)?
ch 3
61
ch 3
62
• How do you detect there are two
separate dots (lights)?
ch 3
63
• Rods are bigger than cones
• Convergence:
ch 3
64
Lateral Inhibition & Mach bands
ch 3
65
ch 3
66
Herman grid
Physical quantities
Perceived quantities
10, 5, 10, 5, 10, 5, 10,
10, 10, 10, 10, 10, 10, 10,
Something is going on inside your brain.
What is it?
Can you explain “something” in terms of the
way neurons talk to each other?
ch 3
67
Something is going on inside your brain.
What is it?
Can you explain “something” in terms of
the way neurons talk to each other?
Some neurons send excitatory (+) signals
(transmitter), and other neurons send inhibitory
(-) signals.
ch 3
68
ch 3
69
ch 3
70
The frequency of
action potential
0
0
The number of
action potential
5 units
emitted by a
neuron is
correlated with
t
Time
the intensity of
the stimulus.
10 units
t
Time
20 units
ch 2
0
71
t
Time
Questions: What happens to B?
0
ch 3
t
72
Questions: What happens to B?
Excitatory
Inhibitory
ch 3
73
Receptive field
• The receptive field of a neuron in the visual
system is the area on the retina that
influences the firing rate (action potential) of
the neuron.
• Measuring the receptive field of a ganglion
cell
ch 3
74
Receptive field of a ganglion cell
Measuring the
frequency of action
potentials elicited by
this ganglion cell.
Cones
Ganglion
cell
B
ch 3
75
Receptive field of a ganglion cell
Ganglion
cell
Cones
12 3 4 5 6 7
B
Firing rate
of B
ch 3
4
3-5
2-6 2-7 76
Questions: What happens to B?
ch 3
77
Measuring the receptive field of a
ganglion cell
Change
the size of the stimulus and78 see
ch 3
the way a ganglion cell respond
Cones
Ganglion
cell
12 3 4 5 6 7
B
ch 3
79
Excitatorycenterinhibitorysurround
receptive
field
Excitatory
Inhibitory
ch 3
80
Questions: What happens to B?
Excitatory
Inhibitory
ch 3
81
Excitatory and inhibitory
connections
• What neurons transmit is electricity.
• Some neurons send positive (excitatory)
signals (+)  increase the firing rate of the
target neuron.
• some neurons send negative (inhibitory)
signals (-)  depress the firing rate of the
target neuron.
ch 3
82
Spatial Summation
a1
+
a2
+
a3 a4
+
+
B
=4
a1
a2
The firing rate
+
+
of neuron B
can be
expressed by
the overall
summation of
the signals that
B receives.
ch 3
c1 c2 c3 c4
a3 a4
+
+
-
-
-
-
B
=0
83
ch 3
84
ch 3
85
How does this happen?
ch 3
86
=sum(B
=sum(B)
ch 3
87
ch 3
88
Fig. 3-6, p. 50
ch 3
89
Fig. 3-7, p. 51
Why is this important?
•  help you to detect the edge of a figure
ch 3
90
ch 3
91
Light
intensity
Perceived
Light intensity
location
location
ch 3
92
Physical stimuli
ch 3
Your perception
93
Lateral
inhibition
ch 3
94
100
20
-- + + ---- + + --- + + ---- + + --
ch 3
95
ch 3
96
ch 3
97
White’s illusion
• Can you explain this by lateral inhibition?
ch 3
98
ch 3
99
ch 3
100
ch 3
101
ch 3
102
ch 3
103
ch 3
104
ch 3
105
Application: Machine vision
• Implementing lateral inhibition into a
computer program.
ch 3
106
Image
ch 3
107
ch 3
108
ch 3
109
ch 3
110
• Edge detection algorithm
– Zero-crossing
ch 3
111