LAB 7 (CHAPTER 10): OBJECT PERCEPTION
(graded out of 50, worth 5% of final grade)
This lab consists of a two parts, which may be done in the lab or at home. Neither
part requires the assistance of a partner.
Equipment:
No special equipment is needed for this lab, just a copy of this hand-out, and a pencil.
Handing In:
Course website only, no other submission methods accepted. Failure to follow this
instruction properly will lead to a 10% mark deduction.
General Background: In class we discussed how early visual processing ultimately
leads to our ability to perceive and interact with objects. Here we will explore two
facets of this topic. One is multistable figures. These were some of the earliest and
greatest challenges to Wundt's structuralist approach to object perception. The
second concept we will explore is modern structural theories. In particular, we will
examine Biederman's "Recognition By Components" model of object recognition, a
highly influential and contentious theory about how the brain models objects in the
world. Prior to doing this exercise, you should read Chapter 10, esp. section C2.
Part 1: Qualitative Observation of a Multistable Figure
One way that researchers study perceptual phenomena is that they simply observe
them. In the first part of this lab we will practice that method, sometimes referred
to a qualitative observation (a.k.a., phenomenology, naturalistic observation).
Initial Observations: As a first step, take a look at the figure on the preceding page.
Just look at it for a minute or two and record what you see (use objective, scientific
language and include technical terms you’ve learned):
Having seen this optical illusion before in a previous psychology class, I knew it was
a cube with certain lines missing, making the whole picture act as somewhat of an
optical illusion. After looking at the shape for a long time, I realized that it was white
background composed of lots of black circles with white lines, which when moved
together, gave an impression of a partially apparent cube shape. The principle of
closure can also be applied to this patterns as when the black circles were moved
closer together they made a cube shape, however, when imagined far apart from
each other, they just looked like an abstract black geometric pattern of circles on a
white background. The law of simplicity can also be present here as one would
immediately try to organize the circles in order to find the most concrete or
apparent shape, in this case, probably being a 3D cube. The law of fate can also be
applied here since we see the circles as moving in uniform parallel/diagonal lines,
instead of being mere circles placed against a white background. The Gestalt pattern
can also be applied to this shape as, as the cube does not actually exist in the actual
image, but it is the human mind attempting to piece apart the illusion and create a
common shape that one has perceived before in past experiences, in this case, one
would probably try to fill in the spaces between the circles in order to create a 3D
cube like shape. I also used the law of proximity while I was analyzing the shape as
when I looked at the closely packed circles, I could see a cube like figure. However,
when I looked at one circle in the top right corner, the image immediately shifted
perspectives as the circles appeared to spread further apart from each other. The
law of similarity can also be applied to this analysis as we see the cube figure due to
the similar circles sharing common features in shape, colour, and brightness; each
circle having a corresponding pair or matching circle in the image.
Some Questions to Consider: Now that you’ve had some time to examine the
figure in an open-ended way, consider the following:
The figure is 2D, but people typically report seeing a cube, which is a 3D structure.
In fact, there are at least 4 versions of a cube that people commonly report. That is,
people typically see the cube either floating in front of eight black circles, or being
seen behind a white wall with eight holes. For either of these interpretations, one
can see the cube as going up and to the right or down and to the left, like so:
This is what is called a “multistable” figure, in that one’s perceptions can stabilize on
one of several interpretations. This is despite the fact that the “sensations” on the
retina remain the same.
Another interesting thing about the figure is that the lines that connect the corners
of the cubes aren’t actually there. They’re only “implied” by the corners, yet people
still perceive them. For some, the lines merely “seem to be there”, while others
report actually seeing faint illusory lines.
Now that you’ve read this over, take another look at the figure on page 1 before
answering the questions below.
Further Observations
1. How many different three-dimensional interpretations of the figure did you see?
Although the figure-ground segregation phenomena in the Gestalt theory states that
based on observation, a foreground impression or a strongly apparent pattern tends
to stand out and is highly distinguishable from any background stimulation, making
many usually see one significant image at first, in this case it would probably be a
cube shape made out of white lines. When analyzing the picture from a few
angles/perspectives, I saw that there were three different interpretations that really
caught my eye. The first time I looked at the figure, I tried to logically make a
common shape out of the circles by piecing the invisible lines together to create a
normal cube with 12 edges and 8 points. The second time that I attempted to look
at the figure, I tried to look outside the supposed cube’s lines and found four pairs of
black circles with white lines crossing through them. The third time that I looked at
the figure with my head at an angle, I saw that the circles created a 2D hexagonal
polygon on top of a series of black circles.
2. If you saw multiple interpretations, did your perception of them shift back and
forth on its own? Could you make it shift? Were there any tricks you used to make
it shift?
I felt that in order to make the figure appear in a certain manner instead of an
unclear optical illusion (which is what I saw when I hazily or unconsciously gazed
upon the figure), I had to consciously follow certain contours and different angles in
order to make the shape shift. In order to see the cube shape I tried to bring my eyes
closer towards the center of the image, which simultaneously made the dots
unconsciously move closer in my mind, while I tried to differentiate the black
shade/circles from the white contrast, in order to make out the shape of a cube. I
could make my perception shift back and forth by trying to look for specific shapes
in the pattern instead of a cube shape, by paying attention to the outside of the cube
shape which immediately made me switch mindsets and see 4 pairs of circles with
different patterns of lines in them. I could easily turn the supposed cube shape into a
2D hexagon composed of many different polygons by looking at the picture with a
tilted head, which at a specific angle, immediately erased the perception of a 3D
cube, leaving me with a whole load of trapezoids, squares and triangles, instead of
black circles that are partially cut with white lines in order to make out a
distinguishable 3D cube shape.
3. The lines that connect the corners of the cube aren’t actually drawn, but one
perceives them. How strong was your impression of the lines? Did you feel like you
could actually “see” them, or was there merely a “sense that they were there”? Did
the strength of the perception vary depending on whether you saw the cube floating
out front vs. behind?
Initially, my impression of the lines was actually quite strong as I used the gestalt
theory to fill in the void spaces, which would most probably be represented by
intersecting the line shapes within the black circles in order to make a 3D cube
shape. Relating to the Gestaltian phenomena of holisticism (or the need to create a
“whole” instead of separate parts), I initially looked at the shape in order to find a
pattern that was rather prominent, in this case, trying to form a geometric pattern
out of the lines in the black circles, in my perspective, the most prominent figure
was a cube shape. Relating to the Gestaltian theory of figure- ground segregation,
the cube was nothing but a subjective impression that tends to stick out from
background stimulation when we try and look for a particular shape (in this case we
are looking for a common pattern, i.e. a cube, and perhaps our consciousness can
make us focus on finding a distinct shape instead of making us solely see a
meaningless cluster of patterned circles). I felt as if I could initially sense that the
corners/edges of a cube were actually there as I tend to look at things in a “holistic
manner” or as a “whole” figure. The strength of the perception actually did vary as it
seemed stronger when I viewed at the front of the shape (as it is presented at
somewhat of a diagonal angle), than when I looked at the back of the shape.
4. Please note here any other observation(s) you might have had about the stimulus:
One other observation that I had about the stimulus, is that the shape tends to shift
with the intensity of one’s visual focus. As I opened my eyes wider I began to see a
widespread image of 8 circles in a geometric pattern (4 dots sandwiched in between
2 rows of two dots), and when I focused my eyes very carefully by squinting/moving
them closer towards the center, I immediately pieced together the outline of a cube
shape in between the circles.
5. Did your observations change after you read the background material above?
How? Why?
My observations did change somewhat after I read that these lines were not actually
drawn in, but it is truly our brain and various cortical processing theories that
subjectively affect the way we perceive objects. Initially, I thought that the cube
shape was drawn in behind some circles in order to test different/subjective views
of vision as well as to strategically test if one is vigilant enough to notice the cube
figure or not. After reading that this figure is a perception test, I began to approach
the figure from different Gestaltian perspectives (e.g. testing the law of similarity as
well as the principle of subjective contours which states that one can subjectively
see edges and distinct shapes where there is no physical luminance, colour or
texture difference), and eventually came to understand that vision and optical
images are based off of one’s personal perception instead of generic uniform
thoughts about visual phenomena.
Instructions for Lab 4, Part 2:
In class we discussed Biederman’s RBC structural description model of object
recognition. In this part of the lab we will look at those models using a naturalistic
observation approach.
Data Gathering
Pick two objects arbitrarily from your environment. These can be anything, natural
or artificial, big or small, animate or not.
1. Sketch out the geon structural description of the two objects below. That is,
sketch out the object as a set of geons such as cubes, cones, spheres, bent
cylinders, etc (see page 346 of the text). Don’t worry, this is not an art class,
but do try to keep things tidy. Label each of the objects and each of its geons.
Bent cylindrical figure
Thin
rectangular
prism
shape
Cylinder
Rectangular Prism
Coffee Mug
Cash Box
Bent cylindrical figure
All of the
separate Geons
2. How easy or hard was it to break the two objects down into geons? Did this differ
between the two objects? Why?
It was rather easy to break both my cash box and my coffee mug down into their
primitive geons or basic shapes. As the vessel of the mug is nevertheless a normal
stout cylinder with a semicircular/bent cylindrical handle, I immediately broke it
down into its basic two geons or geometric shapes. However, for my cash box, it was
slightly harder as the box contained more certain 3D shapes on one of its faces
(represented by the cylindrical handle and the prominent rectangle or the lock that
is right under the handle). Previous experience seeing and using a lot of coffee mugs
also aided with my perception and processing speed, as I only occasionally see and
use cash boxes, compared to the amount of time I spend using and seeing mugs in
everyday life.
3. Do you think the two objects could be differentiated based on their geon
structural description? Could they be recognized from the geons alone? Why or
why not?
Since the cash box is rather similar to a normal rectangular prism shaped box with a
lid on it, and a distinct line (representing the two halves of the box) that runs around
it, one can probably distinguish that it is a container used for holding something. In
most cases as people would observe the geonic prism shape alongside its cylindrical
handle on top of a thin rectangular prism (the common shape of a keepsake/money
box closure/lock) it can be easily implied that it is a cash box, and can somewhat be
easily recognized by its distinct geons. However, when one sees the mug with its
distinct geometric/cylindrical components, they can easily label and differentiate it
from another item, signifying that it can be recognized solely by its geons.
4. Based on their geons alone, would the two objects both be recognizable from
various angles?
The coffee mug is rather recognizable from the majority of angles since it is a rather
generic shape, composed of a short rotund cylinder connected by a thinner curved
cylindrical shape for a handle. View invariance, or the angle and altitude that the
object is looked at, can also have a profound effect on one’s recognition
process/object differentiation time. From a horizontal distance at eye level, if the
round handle of the coffee mug is not showing alongside the cylinder, one may have
trouble differentiating it from a similar object (e.g. a cylindrical coffee canister). One
can notice that the coffee mug is a hollow cylindrical shape with a bent cylinder on
its side for a handle, heavily signifying (based on geonic properties/personal
opinions due to previous experience), that this is infact a coffee/drinking mug.
However, the cash box can easily be pieced apart into geonic shapes and recognized
to be a container at different angles (e.g. when viewed at an elevated height and
from a diagonal perspective). The easiest angle or perspective of view invariance
used to clearly piece apart the all of the box’s geonic shapes would be when looking
at it from the front as one can notice the specific geonic shapes of the cylindrical
handle and the rectangular lock base, signifying that it is most likely a container for
money/keepsakes.
5. Based on your observations here, and your personal experience with object
recognition, what are your thoughts about Biederman’s geon structural description
model?
I actually agree that Biederman’s geon model is frequently used when trying to
differentiate objects from one another and to piece apart the exact shape of certain
objects (e.g. a briefcase which is usually composed of two geons, a semicircular
cylinder which is connected to a prism shaped box). Whenever I approach a new
object, I try to identify it in a top down process, which is similar to Biederman’s
theory, as I take the whole object and break it down into smaller shapes, in order to
aid with the recognition process. View invariance is also another key factor in my
mental processes, as I view and try and identify the whole object from different
angles (e.g., looking at the cup from the front where its handle cannot be seen as it
is usually on one side of the mug, then turning it around to see if there is a handle or
cylindrical figure on the side when looking from a different angle, I also to see if the
cylinder is hollow [marking if it is a mug or not] by looking at it from an elevated
angle), before I make up my mind and come to a conclusion what something really is.
In this lab, I found Biederman’s model to be quite effective in recognizing/recalling
what my chosen objects were, even though I possess them and have used them
multiple times in the past. Breaking things down into smaller components, (in this
case 3D shapes), may seem like a rather slow and unnecessary process; however, it
initially helps clarify and recall past memories about similar items used and their
specific purposes, making one have a higher chance of identifying the item correctly.