object perception (ch 5) - exam 2 (2/3)

lower level processing

visual processing that includes seeing without processing what an item is - identifying local elements

higher level processing

visual processing that includes interpreting and making meaning from visual input beyond its basic images/parts - objects and form

• detection, discrimination, identification

detection, discrimination, identification

3 parts of higher level processing

detection

part of HL processing - picking up on an object’s presence

discrimination

part of HL processing - distinguishing an object from other objects

identification

part of HL processing - recognizing an object for what it is from memory

structuralist and Gestalt

approaches to form perception

structuralist

approach to form perception - concerned with the parts/elements of a form, individual

• Biderman’s recognition-by-components

• Treisman’s feature integration theory

gestalt

approach to form perception - concerned with how the parts are combined

• holistic approach to the brain that prioritizes HL processing over LL processing

• we see the scene, rather than the scene’s parts

• stimuli has enough meaning to be directly interpreted

• 4 laws: proximity, similarity, good continuation, closure

proximity, similarity, good continuation, closure

gestalt principles of organization/grouping

proximity

gestalt principle of organization/grouping that states that items near each other are grouped together

similarity

gestalt principle of organization/grouping that states that items similar to (look like) each otherare grouped together

good continuation

gestalt principle of organization/grouping that states that we see items intersecting, aligned, or overlapping as continuous, rather than as broken fragments

• smooth edges are more likely to be seen as continuous than those with sharp/abrupt edges

closure

gestalt principle of organization/grouping that states that we fill in missing parts/gaps of an incomplete shape to perceive it as unified/complete

feature integration theory (Treisman)

theory about object construction

• objects are constructed from their elements, or a set of primitive parts, that are put together to be perceived as a whole

  • bottom up

  • 2D

  • features at particular locations are combined, forming an object in a particular location for recognition

• three stages: pre-attentive, focused attention, recognition

• determining basic features: pop-out boundary, visual search

• evidence: illusory conjunctions

pre-attentive, focused attention, recognition

stages of feature integration theory

pre-attentive stage

first stage of feature integration theory: quickly ID basic parts/features without a lot of attention (unconscious)

focused attention stage

second stage of feature integration theory: basic features are combined to form an object

recognition

third stage of feature integration theory: comparing object with memory to recognize it for what it is

pop-out boundary

strategy to determine basic features in FIT - displayed among the noise to make an object obvious

• ex: circle made of / in a sea of \

visual search

strategy to determine basic features in FIT - objects with unique features are easier to find

• ex: O in a sea of V’s is easy to find, but finding an O in a sea of P’s, Q’s, R’s, and V’s is significantly harder

  • roundness no longer unique

illusory conjunction

evidence to support FIT - incorrect combination of features

• flashing colored letters then having people repeat what they saw - correct features, but wrong letters = wrong combination

• demonstrates that objects are made up of individual features

recognition by components (Biederman)

theory about object constuction that objects are made up of approximately 36 different geons/figures (simple shapes)

• principle of componential recovery

• geon properties: view invariance, discriminability, resistance to visual noise

  • harder to recognize when visual noise blocks where 2 geons intersect

principle of componential recovery

principle in RBC - if enough geons of an object are present, we can recognize the object

view invariance

property of geon in RBC - recognizable at any viewpoint, orientation, and/or angle

discriminability

property of geon in RBC - distunguishable from 1 another (e.g., cylinder v. pyramid)

resistance to visual noise

property of geon in RBC - can be seen, viewed, and understood amongst visual noise

view invariance, discriminability, resistance to visual noise

properties of geons (RBC)

we can recognize broad classes of objects, but does not explain why we can recognize different objects within a category

critique of RBC

IT cortex (inferotemporal)

cells in this area allow us to recognize faces as they prefer face stimuli

specialized cells in IT cortex, ability to distinguish, expertise, provides info (age, gender, emotion, etc.)

why are faces special?

normal face perception and prosopagnosia

evidence of individualized processing of faces comes from studies of…

prosopagnosia

inability to recognize faces (okay with other objects)

face inversion effect

faces are hard to recognize when they are upside-down, but objects aren’t

• demonstrates that faces are processed differently than objects

thatcher effect

upside-down face with mouth and eyes flipped upside down looks normal, but once it’s placed right-side up, it looks deformed/wrong

Treisman

pioneer of feature integration theory

Biederman

pioneer of recognition-by-components