Someone recently bought our

students are currently browsing our notes.

X

4 Object Perception Notes

Psychology Notes > Perception (2nd year) Notes

This is an extract of our 4 Object Perception document, which we sell as part of our Perception (2nd year) Notes collection written by the top tier of Durham University students.

The following is a more accessble plain text extract of the PDF sample above, taken from our Perception (2nd year) Notes. Due to the challenges of extracting text from PDFs, it will have odd formatting:

Object perception
Ventral Stream
Ventral stream = object and face recognition
Starts from V1 and includes the extrastriate visual areas V2, V4 and IT
V1 is important for vision and has some simple 'feature detectors' BUT not sufficient for object recognition:

1. Small RFs of V1 neurons  only 'see' a small portion of an object

2. V1 neurons only sensitive to simple features eg short line segments of diff orientation = not enough to separate a face from a house
V2 have more complex preferences than V1:
 Slightly large RFs
 Interesting response properties

May respond to illusory boundaries for which V1 neurons are blind

Might know whether an edge belongs to an object or is part of the background
= important response properties but still not sufficient
V4 and IT cortex neurons respond to even more complex forms:
Have larger RFs  more invariant to object location in the VF
Ie respond to a preferred stimulus anywhere in their large RFs
V1
V2
V4
IT

Lines, edges
Curves, textures
More complex shapes
Shapes, objects

Faces

Tanaka, 1996:
IT neurons responds best to an 'apple' shape: circular body with a bar sticking out form lower right
Other IT neurons prefer other complex shapes
Bruce, Desimone & Gross, 1981:
Some neurons respond preferentially to face stimuli
Desimone et al., 1983: some cells respond best to certain profiles of faces = view-specific cells
 others show view-invariance

Objects are put together at higher levels of processing than V1 where RFs are larger and have more complex properties
 RFs increase in size from V1 to higher visual areas
Perceptual constancy
= how we recognise an object from different viewpoints and distances
 important concept in object recognition
Viewpoint, distance and illumination = all incidental features that are not important for object identity

Illusory form: Kanisza
How do we combine small features into whole objects?
Structuralists: we perceive objects from a combination of atoms/basic features
BUT illusory form proves this wrong Object perception
 we view a whole percept even though there is no whole
 We see an arrow although there are no continuous contours
 Black shapes in the corners of the triangle are more likely to be round shapes covered by a white triangle than black pacmen suspended in thin air
= Brain takes into account probabilities  calculates the most likely cause
Ambiguous figures:
The brain usually interprets images based on what is most likely
BUT ambiguous figures show that sometimes our visual system is baffled and cannot arrive at one interpretation  2
Eg face or vase

Gestalt school
= we directly perceive wholes based on grouping laws, rather than combining small details into form
The whole (final percept) is more than the sum of its parts (constituent sensations)
Perception > Sensation
Grouping rules:
Good continuation
Proximity
Similarity
Pragnanz

Things that follow from each other smoothly probably belong together
Close-by things appear grouped
Similar things appear grouped
Of possible options, we perceive the simplest option
Koffka, (1935): of several geometrically possible organisations that one will actually occur which possesses the best, simplest and most stable shape

These results describe how the visual system makes sense of the jumble that the retinal image is
BUT they don't explain how the visual system calculates similarity, proximity etc
= not a mechanistic account

Object Constancy
= how we recognise objects from different viewpoints, sizes and locations on the VF
2 explanations which operate at different levels:

1. Neuron level

2. Computational level
Neuron-level explanation: discusses neuronal response properties
IT neurons are:
 Viewpoint-invariant  respond to faces regardless of face orientation
 Size invariant  respond to objects at different retinal sizes, corresponding to different distances from the viewer
 Location invariant  respond regardless of location
V4 neurons show colour constancy  they respond to surface colours regardless of how they are illuminated

Buy the full version of these notes or essay plans and more in our Perception (2nd year) Notes.