Perception Flashcards
Perception
How external world gets represented in our brain/mind so that we can understand and act upon what’s going on around us
Apperceptive agnosia
Associative agnosia
Unable to name/match/discriminate visually presented objects
- Failure to combine visual info to complete percept (deficits in copying)
Unable to associate visual pattern w/ meaning
- Can combine features into a whole so can copy
3 steps to visual perception
Input/sensation
Basic visual components assembled
Meaning linked to visual input
Senses that measure properties of our own body (interoception):
Propioception
Nociception
Equilibrioception
Location of limbs in space
Pain due to internal bodily damage
Sense of balance
Experience error
Inverse projection problem
False assumption that structure of world is directly given from our senses
Perception only uses hints to retrieve 3D object; We only see 2D projection of it
Fixation-saccade cycles
Vision is combo of
- Smooth pursuit movement (when eyes following object; info processed)
- Saccade (eyes shift between scenes; input not processed)
Perception fills in the gaps
Bottom-up processing
Top-down processing
Data driven
Recognize patterns by analyzing sensory input step by step
Conceptually driven
Perception influenced by prior knowledge, memories, exp
Template theory
We have mental stencil for an array of diff patterns
- Important for computers perceiving letters
Not good for humans bcuz everyone has diff writing
Feature matching
Pandemonium theory
Feature detector neurons
We have a system for analyzing each distinct feature of a visual item
Break down into distinct visual features and out them back together
- Whichever demon is shouting loudest is correct
- Serial (between demon types) and parallel (each demon working at same time) processes
**Insufficient bcuz unknown how pieces are put back together
Found in primary visual cortex; responds to specific input/stimulus
Prosopagnosia
Semantic agnosia
Fusiform face area (FFA)
Difficulty recognizing faces
Difficulty recognizing objects except for faces
Region in inferior temporal cortex that shows greatest activity when performing face-specific tasks
Visual streams:
Ventral stream
Dorsal stream
Terminates temporal lobe
- Concerned w/ processing “what”
Terminates parietal lobe
- Concerned w/ processing “where”
Division between them is of perception vs action
Biderman’s Recognition by Components (Geon theory)
3D shapes called geons
To identify object, you match to geon
Recognition is impaired when objects viewed from non-canonical viewpoints (unusual angles)
Humans appear to have viewer centred bias
Object recognition faster from familiar viewpoints
Cortical neurons show viewpoint specificity
Scene schemas
(Top-down processing)
Used to help identify objects in familiar enviros
Gestalt psychology
How perception gets organized into meaningful units
- Whole is different than sum of its parts
Gestalt laws for grouping
Law of proximity (Close together, grouped together)
Law of similarity (Similar, grouped together)
Law of common region (Enclosed within same region, grouped together)
Experience
- Things associated together in prior viewings will be grouped together in future
Direct perception (Gibson)
Ambient optic array
Optic flow
Enviro contains all info needed for perception
- Uses action
Structure imposed in light by enviro, contains info needed
Motion/flow in optic array gives info:
- Flow = Observer is in motion
- Direction of flow = Direction observer is moving in
- Flow coming out of point = Object moving closer
- Flow moving toward point = Perceiver moving away
Objects’ affordances
Goal of perception is to provide perceiver with this info
- Tells us what you can do with object
- Depends on object and observer
Modern researchers’ belief on perception
(Related to Gibson’s view)
Both actions and representations involved in perception
- Action influences how we perceive the world
Ideomotor apraxia
Can’t act out actions w/ objects (how to us)
- Damage to parietal lobe (“where” pathway)
Motor plan
Mirror neurons
Voluntary movement plans goal of action and how it’ll be accomplished
Involved in planning
- Responds equally when performing and viewing an action
Tanaka and Farah
(Face processing)
Found that it’s easier to recognize parts of a house than parts of faces
Face inversion effect
We’re faster and more accurate at recognizing upright faces compared to inverted faces
Diamond and Carey dog identification study
Dog experts perceived dogs similarly to how we perceive faces
(Better upright)
Phonemic restoration effect
Missing sounds are “filled in” by brain based on knowledge of language
Figure-ground assignment
The determination of which side of a boundary contains the shape vs background
General recognition
Ability for a computer to classify a broad class of diff objects
Scene schema
Learned representation of which objects tend to appear in specific kinds of scenes
Convolutional neural networks (CNNs)
Learns features that can appear in diff locations in an image
- Useful for computers to recognize image as being part if category
Kernels or filters: Array of numbers used to detect presence of image features
- Produces feature map (shows how much of the feature is present across diff locations of the image)
