CCC- Week 2- Object recognition

Question 1

What is perception?

Answer 1

Perception refers to our ability to extract meaning from sensory input.

It includes the 5 aristotelian senses audition, taste, touch and olfaction (smell), but research is dominated by vision.

Question 2

How many senses are there estimated to be?

Answer 2

9-22

Question 3

What is proprioception?

Answer 3

Is the sense that lets us perceive the location, movement, and action of parts of the body.
It encompasses a complex of sensations, including perception of joint position and movement, muscle force, and effort.

Question 4

What are the different types of receptors & where are they?

Answer 4

Chemical receptors- drive out taste & smell
Photoreceptors- found in eyes
Mechanoreceptors- Cover lots of things such as touch, like proprioception- muscles etc.
Thermoreceptors- Temperature

Question 5

What are all of our senses reliant on?

Answer 5

One of the 4 receptors.

Question 6

What does vision account for in our cortex?

Answer 6

Over 50% of all the neurons- so lots of research is focused on vision (especially as it is our dominant sense)

Question 7

What sort of process is perception?

Answer 7

A constructive process- not a passive one.

When you look outside, you are not just passively receiving the information & interpreting it- you are actually doing so much more - essentially hallucinating- controlled hallucinations. A lot of what you think you have seen is completely illusive.

Question 8

What is an example of us experiencing an illusion?

Answer 8

All aware we have a blind spot (the point of entry of the optic nerve on the retina insensitive to light- an area where view is obstructed)- it is about the size of an orange at arms length.
When you have both eyes open, your blindspots are each in different places so you don’t see your blindspot. However when you close one eye, you still can’t see a hole in your vision. Your brain fills in that gap- yet you have no sense of their being a gap in your vision (you think all the right details/ color is all there)However it isn’t there - we have a hole in our vision that we cant perceive- yet we perceive it as complete. That is an example of how it is an illusion- we think we see more than we actually do.

A lot of our thinking about perception & our ability is contributed to the eye- yet in reality- the eye is very lousy - what’s impressive is the processing in V1- our visual system- that incorporates what’s coming in ( the visual stimuli) & all of your history & experience & biases- which are suitable for our particular history/ evolutionary history- so theres lots more going on than whats just happening in the eye.

Question 9

What is a blind spot?

Answer 9

The point of entry of the optic nerve on the retina insensitive to light- an area where view is obstructed.

Question 10

The visual system model

Whats in it?

Answer 10

Brain receives sensory input allowing us to perceive an object.

Cognitive system constructs perception.

Question 11

How many neurons in our cortex does vision account for?

Answer 11

50%

Question 12

Processing streams- ventral/ dorsal

What is the dorsal pathway used for?

Answer 12

Spatial integration & location information.

Question 13

What is the ventral pathway used for?

Answer 13

Used for identifying objects (it goes down the temporal cortex)

Question 14

What are the two visual pathways?

Answer 14

Dorsal pathway

Ventral pathway

Question 15

What is the different between the ventral and dorsal pathways?

Answer 15

Ventral- processes visual information- for perception.

Dorsal- processes visual information for the purpose of executing movements.

Question 16

Why is our general belief about how the eye works in the model unrealistic?

Answer 16

We have different visual fields- in each eye- there is a left and right visual field.
The information goes to the two halves and gets processes separately.
Half the information goes to one hemisphere- half goes to the other hemisphere.
In order to see & recognize objects- you have lots of inter hemispheric communication- but we are not aware of this.

Question 17

What happens with individuals who suffer from epilepsy?

Answer 17

Many have a corpus colostomy- surgery to treat epileptic seizures.
Involves cutting the corpus callosum- which stops the brain from sending seizure signals from one half of the brain to the other.
After the surgery- the seizures tend to be less severe- as they will only affect half of the brain.

Question 18

What happens to the image on the retina?

Answer 18

Its inverted!

Question 19

Where does language tend to be specialized?

Answer 19

On the left!

Question 20

The visual system- A misleading impression of simplicity

What study was conducted & on what animals?

What did the study support?

Answer 20

Tootel et al 1982

Study tested on monkeys.
Monkeys- unconscious with eyes propped open.
Were presented with stimulus (25-30 mins) & injected with a radioactive isotope in glucose- so where brain was most active- the isotope got concentrated.

Supported the notion that a near perfect representation of the external world is “projected” onto our primary visual cortex.

Question 21

Object recognition- what is the model called?

Answer 21

Three stage model

Question 22

What is in the Three Stage Model?

Answer 22

Image

Stage 1- Local Features- Edge detection/ contrast

Stage 2- Shape representation- Gestalt principles/ feature
integration

Stage 3- Object representation- Stored representations/
knowledge

Question 23

Three stage model

What happens at stage 1?

Answer 23

Local features- edge detection/ contrast

You’re extracting the local features- simple features- just like lines & edges & contrasts.

Question 24

Three stage model

What happens at stage 2?

Answer 24

Shape representation- Gestalt principles/ feature integration

The basic features are being combined using things like gestalt principles & feature integration to perform larger but still primitive blocks and regions. You have slightly bigger blocks and regions performed from the really basic ones.

Question 25

Three stage model

What happens at stage 3?

Answer 25

Object representation- stored representations/ knowledge

Those are used to map against the object representation- & store representation in memory

Question 26

Gestalt principles

What are these?

Answer 26

, are rules of the organization of perceptual scenes.

The whole visual percept is more than just a sum of its parts.
Our perceptual system tried to impose organisation into its input.
Components of an image are grouped together on the basis of certain visual properties.
Laws of perceptual organisation e/g/ “good continuation” & “closure” give rise to “illusory contours”

Question 27

What are some examples of Gestalt principles?

Answer 27

Proximity
Similarity
Good continuation
Closure

Question 28

What happens in each of the Gestalt principles?

Answer 28

Proximity- another important group. In the image above, it looks like 4 columns. However there is no reason why the first 3 lines might be one object, and the next 3 another object. Your brain doesn’t do this automatically- it uses proximity.

We also use similarity to group- e.g. in the image- it tends to be seen as a row of 0s and and a row of xs etc instead of columns of 0 x 0 x- its just that we use that perception.

Good continuation: in the image- we tend to see it as a straight line dissecting the curved line. However it doesn’t have to be like that- it could be a straight line that then curved- coincidentally- this coincidence doesn’t tend to happen in the natural world- so it makes sense that the line was straight and the other line was curved.

This is the same as it is with closure. We tend to see the image as a circle in front of another circle- than one circle and another circle with a part missing. We see it like this as that is what we would see in the natural environment. (need to know the examples)!

Question 29

What are illusory contours?

Answer 29

A visual illusion, in which people see edges even though they are not physically present in the stimulus (aka subjective contours)

Question 30

Shape perception

What was the idea and model by Marr 1982?

Answer 30

Idea- primarily “bottom up” processes produce a “primal sketch”.
The sketch contains “primitives”- edges/ orientations/ positions/ lengths/ colours etc.
“Top down” processes (e.g. Gestalt laws) are used to group collections of primitives together into lines, curves, larger blobs, groups & small patches”- symbolic primitives.

Question 31

Object recognition

What are the 3 models?

Answer 31

Template matching (prototype theory)
Feature analysis
Recognition by components (structural theory)

Question 32

What is template matching?

Answer 32

A template is an internal representation.
A memory against which the visual input is matched.
Computer based object recognition’s use templates.
Object recognition must involve some kind of contact with a “comparable internal form”

Question 33

What is feature analysis?

Answer 33

• Assumes lower level features are analysed first.
• The perceptual system searches for simple but characteristic
  features of an object.
• Supported by neurological evidence (e.g Orientation selective
  cells in the visual cortex)
• Most research focuses on letter/ word recognition (used to
  read postcodes)
• Letter A made up of /l-

Question 34

Recognition by Components (Feature analysis in 3d)

What did Biederman 1987 state?

Answer 34

Any specific view of an object can be represented as an arrangement of simple 3D shapes - geons.

Question 35

Recognition by Components (Feature analysis in 3d)

What are geons?

Answer 35

Geons are the simple 2D or 3D forms such as cylinders, bricks, wedges, cones, circles and rectangles corresponding to the simple parts of an object in Biederman’s recognition-by-components theory.

The theory proposes that the visual input is matched against structural representations of objects in the brain.

Question 36

Recognition by Components (Feature analysis in 3d)

Geons are viewpoint invariant- what does this mean?

Answer 36

Enables us to recognize objects regardless of viewing angle; this is known as viewpoint invariance.

They are easily recoverable from a 2D retinal image

Question 37

Recognition by Components (Feature analysis in 3d)

What invariant properties are there?

Answer 37

Cotermination
Parallelism

Question 38

Recognition by Components (Feature analysis in 3d)

When is object recognition impaired?

Answer 38

When geons are made non-recoverable by removing termination points.