Chapter 5 Flashcards
What is computer vision? (Object recognition)
Detecting objects based on stored representations
Ex. Vehicles, cell phones (face ID)
Problem with computer vision?
Often struggle with vague/ degraded/ odd images and scenes
What is the inverse projection problem?
Image on a retina can be any object
Computers often fall for this
We don’t cuz our brain interprets what we see
What is viewpoint invariance?
The ability to recognize an object regardless of the viewpoint
Hard for computers
What is perceptual organizations? (Wundt)
Perception is combining sensations
Grouping visual elements together (organization) so that you can determine the meaning of a whole (perception)
Can structuralism explain apparent movement?
No
Explain the three Gestalt approaches to perceptual grouping (structuralism, apparent movement, illusory contours)
Structuralism: distinguished between sensations and perception
Apparent movement: illusion of movement
Illusory contours: look real but don’t have physical edge (weird triangle illusion)
Explain the Gestalt approach to grouping (Good continuation)
Connected points resulting in straight / smooth curves belonging together
Explain the Gestalt approach to grouping (Pragnz)
Each stimulus is seen as simply as possible
Explain the Gestalt approach to grouping (Similarity)
Similar things are grouped together
Explain the Gestalt approach to grouping (Proximity)
Close things are grouped
Explain the Gestalt approach to grouping (Common fate)
Things moving in the same direction are grouped
Explain the Gestalt approach to grouping (Common Region)
Elements in same region are grouped
Explain the Gestalt approach to grouping (Uniform connectedness)
Connected region of visual properties are perceived as a single unit
What is figure ground segregation?
What is figure and what is background
What are some of the properties of the figure and ground?
-figure is thinglike
-figure in front
-figure has contour/ lineart
-figure is usually in lower area
-ground goes behind
Explain how we determine which is the figure? (Meaningfulness and recognition by components)
Meaningfulness: if figure is meaningful it’s easier to see (human silhouette)
Recognition by components: objects are split in basic shapes (geons)
Issues:
-doesn’t explain grouping
-doesn’t count distinguishing mugs
What is a scene?
- Background elements (acted within)
- Objects in meaningful way (acted upon)
How long to perceive the gist of a scene?
250 Ms
90% accuracy
How can we Perceive the gist of a scene so quickly?
-naturalness
-openness
-roughness
-expansion
-colour
How do we use properties of the environment to identify scenes?
Blue usually equals open sky / ocean
Green usually equals landscapes
Vertical / horizontals usually equal buildings
Explain physical regularities?
Regularly occurring physical properties of the environment (natural and human made)
What are semantic regularities and scene schemas?
You’d expect a knife in the kitchen not a car
What is the likelihood principle?
Perceive and object that us most likely to have caused the stimulation
