Sensation And Perception Rutgers 2018 (MgGann) Exam 2

The problem of object recognition is the pictures were just a bunch of..

pixels on a screen, but in each case you perceived a house

Object recognition happens in as little as

150 ms.

Because object recognition is so fast, there cannot be ....

a lot of feedback from higher brain areas

Mid-level vision is a loosely defined stage of

of visual processing that comes after basic features have been extracted from the image (low-level vision) and before object recognition and scene understanding (high-level vision).

Mid-Level Vision involves the perception

of edges and surfaces

Mid-level Vision determines which regions of an...

image should be grouped together into objects

How do you find the edges of objects?

Cells in primary visual cortex have small receptive fields

Computer-based edge detectors are ________as good as humans; sometimes computers don't find enough _______ or they find ...

Not, edges, too many edges

Illusory contour

A contour that is perceived even though nothing changes from one side of the contour to the other in the image

Gestalt: In German means...

"form" or "whole"

Gestalt psychology:

The whole is greater than the sum of its parts."

structuralism, that emphasizes...

the basic elements of perception

Gestalt grouping rules is...

A set of rules that describe when elements in an image will appear to group together

Good continuation is a A Gestalt grouping rule stating...

that two elements will tend to group together if they lie on the same contour.

in Middle vision, Some contours in an image will group

because of good continuation

Texture segmentation is

Carving an image into regions of common texture properties

Two specific rules of Gestalt grouping rules...

Similarly and Proximity

Parallelism

Parallel contours are likely to belong to the same figure

Symmetry

Symmetrical regions are more likely to be seen as figure

Common region:

Two features will group if they appear to be part of the same larger region

Connectedness:

Two items will tend to group if they are connected

Camouflage: Animals exploit .

Gestalt grouping principles to group into their surroundings

Sometimes camouflage is used

to confuse the observer.

Perceptual committees are

a metaphor for how perception works

Many different and sometimes competing principles are

involved in perception

Perception results

from the consensus that emerges

Committee rules:

Honor physics and avoid accidents

Ambiguous figure is

A visual stimulus that gives rise to two or more interpretations of its identity or structure

Perceptual committees tend to obey

the laws of physics

Accidental viewpoint is

A viewing position that produces some regularity in the visual image that is not present in the world

Perceptual committees assume

viewpoints are not accidental

The wire-frame Necker Cube

Templates versus structural descriptions

Template theory: The proposal that the visual system recognizes objects by matching the neural representation of the image with a stored representation of the same "shape" in the brain.

Structural description: A description of an object in terms of its parts and the relationships between those parts.

Structural description:

A description of an object in terms of its parts and the relationships between those parts.

Template theory:

The proposal that the visual system recognizes objects by matching the neural representation of the image with a stored representation of the same "shape" in the brain.

The pandemonium model was created by

Oliver Selfridge's (1959)

The Pandemonium model is a single model of

letter recognition based on templates of letters

Perceptual committee can be made up of

"demons"

Demons loosely represent

neurons

Each level is a

different brain area

Figure-ground assignment is

The process of determining that some regions of an image belong to a foreground object (figure) and other regions are part of the background (ground)

Gestalt figure-ground assignment principles are

Soundness, size, symmetry, parallelism, relative motion

Surroundedness:

The surrounding region is likely to be ground

Size:

The smaller region is likely to be figure

Parallelism:

Regions with parallel contours tend to be seen as figure

Relative motion:

If one region moves in front of another, then the closer region is figure

Dealing with occlusion,

Relatability is the

degree to which two line segments appear to be part of the same contour

Parts and wholes,

Global Superiority Effect is the properties of the

whole object take precedence over the properties of parts of the object

Nonaccidental feature is a feature of an object that is NOT

dependent on the exact (or accidental) viewing position of the observer

T junctions Indicate

occlusion

The top of the T is in

front

The Stem of the T is in the

back

Y junctions indicate corners facing the

observer

Arrow Junctions indicate corners facing

away from the observer

Goals of mid-level vision:

1. Bring together that which should be brought together

2. Split asunder that which should be split asunder

3. Use what you know

4. Avoid accidents

5. Seek consensus and avoid ambiguity

Neural representation of sensory stimuli is the idea that specific nerual activty patterns are caused by

specific outside stimuli, such that observing the neural activity allows us to infer the outside stimulus.

Naïve template theory is the idea that the visual system recognize objects by

matching the neural representation of the image with a stored representation of the same "shape" in the brain

Example of Naïve template theory: what if a group of V1 neurons that were tuned to angled lines and horizontal lines fed into a single cell that would thus

respond to the letter "A" in that part of the visual field. That would be an A template.

Problems with the Naïve Template theory is

We'd need an almost infinite number of these templates

Templates are not viewpoint-invariant, but object recognition (mostly) is.

Viewpoint invariance:

1. A property of an object that does not change when an observer changes viewpoint

2. A class of theories of object recognition that proposes representations of objects that do not change when viewpoint changes

Structural description theory is The idea that the visual system recognizes objects by

figuring out each object's constituent parts and the relationships between those parts. This permits perfect viewpoint-invariance.

Example of Structural Descirp Theory

Biederman's "recognition by components" model, in which the brain recognizes objects by figuring out what combination of "geons" composes them.

Problems with Structural Descrip Theory are

NOT totally viewpoint invariant.

Newly learned objects are easier to recognize if they haven't been rotated.

Geons aren't always the best descriptions of objects

Geons in Biederman's "recognition by components" model are the

"geometric ions" out of which objects are built

"Greebles"

are artificial shapes used to test object recognition

Viewpoint invariance:

1.

2.

1. A property of an object that does not change when an observer changes viewpoint

2. A class of theories of object recognition that proposes representations of objects that do not change when viewpoint changes

Entry-level category

For an object, the label that comes to mind most quickly when we identify the object (note this is usually a narrower category for experts than for non-experts).

Subordinate-level category

A more specific term for an object

Superordinate-level category

A more general term for an object

Face recognition

seems to be special and different from object recognition

Prosopagnosia

An inability to recognize faces

Agnosia:

A failure to recognize objects despite being able to see them

Inferotemporal (IT) cortex is a part of the

cerebral cortex in the lower portion of the temporal lobe

Inferotemporal (IT) cortex is important for.. It is part of the

object recognition, "what" pathway

Lesion

1. (n.) A region of damaged brain

2. (v.) To destroy a section of the brain

When IT cortex is lesioned, it leads to

agnosia

agnosia

A failure to recognize objects despite being able to see them

Grandmother cells (extreme localization of function)

Could a single neuron be responsible for recognizing your grandmother?

Feed-forward process is a process that carries out a

computation one neural step after another, without need for feedback from a later stage to an earlier stage

Object recognition occurs so quickly that

feed-forward processes must be occurring

Feedforward circuits could in principle quickly spot the

coincidence of features characteristic of your grandmother.

The Problem with Feed-foward process

here seems to be a strong interplay of "top-down" effects of expectations and prior knowledge with "bottom" information about the physical details of the stimulus.

Recorded from

single neurons (single units) or small groups of neurons (multi-units) in the human medial temporal lobe.

The location is in the "

What" pathway for higher level visual processing.

Subjects were 8 patients with severe

pharmacologically-intractable epilepsy who had electrodes implanted to look for the epileptic focus

Patients sat in bed and watched pictures on a laptop in

30 minute sessions while the data from the electrodes was recorded.

Pictures were each presented

6 times, for one second each

The picture covered about

1.5 degrees of visual angle.

About ____ pictures were shown per session

88

Patients had to

press a button after each picture to say if it was of a human or not (to make sure they paid attention).

When they looked at the data afterwards, they counted the

action potentials (spikes) recorded from individual neurons before the picture was displayed (baseline) and after the picture was displayed.

Neurons were considered to have

responded to the picture if they exhibited a response rate at least 5 standard deviations above baseline.

Range of selectivities they found

1.38 neurons responded to specific people (Aniston, Berry, Kobe Bryant, Julia Roberts, etc)

2.6 responded to specific buildings (Tower of Pisa, Sydney Opera House, Baha'i Temple)

3.5 responded to specific animals (spiders, seals, or horses)

4. 2 responded to specific food items

_____ out of 132 neurons exhibited selectivity.

51

8 of these cells responded to both

the picture of the person/place and their name.

Experiment conclusions

1. hese cells must be responding to really abstract representations, because the physical details of the images were so different (line drawings, angles, backgrounds, text, etc).

2. Maybe there's some truth to the idea that there are "grandmother cells" after all.

3. This could be important for how memories are created in the brain.

What are the seven colors of the rainbow?

R.O.Y.G.B.I.V.

The Bullshit Rainbow are

as real as a leprechaun

Is there a wavelength perceived as violet?

Yes

Problem of univariance is an infinite set of different

wavelength-intensity combinations can elicit exactly the same response from a single type of photoreceptor