Midterm 2: Topic 3 Objects and Scenes (Part 1,2,3) Flashcards
the inverse projection problem
2D image on retina from 3D world
Infinite number of objects that can create a given pattern of light on the retina
Infinite pattern of light on the retina
Ambiguous Cylinder Illusion
Object with a mirror behind it
Creates the illusion through the mirror that there is different shape than what is in front of the mirror
Based on perspective
The object is a blend of a square and a circle
Will either look like a square or circle based on how you are looking at it
Brain makes an assumption on the shape and goes with the most probable answer
what is Viewpoint invariance
Ability to recognize an object regardless of viewpoint
what is Gestalt Approach
a theory of perception that emphasizes that the whole is greater than the sum of its parts. It suggests that humans naturally organize sensory information into patterns and wholes, rather than just perceiving isolated elements.
what causes us to group elements together
Principles of perceptual grouping
what is Principle of good continuation
Lines tend to be seen as following the smoothest path
what is Principle of good figure (simplicity or prägnanz)
Stimulus patterns are seen so the resulting structure is as simple as possible
what is Principle of similarity
Similar things appear grouped together
what is Principle of proximity (nearness)
Nearby objects appear grouped together
what is Principle of common fate
Elements that in same pattern together appear grouped together
what is Principle of common region
Elements that are in the same region of space appear to be grouped together
what is Principle of uniform connectedness
A connected region with the same visual properties is perceived as a single unit
What causes us to separate elements apart?
Principles of perceptual segregation
what is figure ground separation
Figure is more “thing-like” and memorable than ground
Figure is seen in front of ground
Ground is more uniform and extends behind figure (the background)
What features help determine what we perceive as foreground and what as background?
whether something’s lower or higher in the scene.
lower = figure
higher = ground
are convex shapes figure or ground?
More likely to perceive something as the figure when there is convex edges
flaw to Gestalt’s approach
Lacks specificity
bottom up - Relying primarily on the stimulus and not on past experience
what is Recognition-by-components (Bottom up processing)
We perceive objects by perceiving elementary features
An object is recognized when enough information is available to identify its’ geons
what are geons
three-dimensional volumes
Geons can be distinguished from other geons from almost all viewpoints
Geons can be perceived in “noisy” conditions
Recognizable no matter the illumination direction, surface markings, and texture.
36 distinct geons
what is Principle of Componential Recovery
Key to object recognition is not amount of information, but ability to identify components (geons)
When more key characteristics (joints and attachments) are visible the easier it is to detect the object
experiments on scene perception suggest what
we can perceive large scale properties first (< 250 ms), and then more slowly fill in details (~500 ms)
How do we perceive gist?
Global image features
Degree of naturalness
Degree of openness
degree of roughness
degree of expansion
colour
We simultaneously process visual scene at multiple spatial scales or frequencies, Low frequency is for what, high frequency is for what?
Low frequencies: gist
General themes related to the whole scene
High frequencies: detail
Edges of objects and things present within the scene
what are hybrid images
Decomposed images into high and low frequencies and then overlapped them with each other
Normally information at low and high spatial frequencies is complementary, but in a hybrid image, it conflicts.
how does experience change perception (top down)
Brain uses context
Meaningfulness/familiarity influencing figure/ground separation
Likelihood principle
We perceive the object most likely to have cause the pattern of stimuli we receive
Unconscious inference
The application of the likelihood principle is unconscious, but based on past experience
It is also quite rapid
Bayesian inference
𝑃 (𝐴|𝐵) = 𝑃(𝐵|𝐴) 𝑃(𝐴)
Refers to how one comes up with and decides on what hypothesis is most likely by combining our prior beliefs and experience with the current evidence in front of us.
𝑃 (𝐴|𝐵) = ?
posterior belief
𝑃 (𝐵|𝐴) =?
likelihood
𝑃(𝐴) = ?
prior belief
objects and scenes have strong association with what stream
ventral what stream of the brain
Starts in V1 and heads forward (rostrally) to the temporal lobe
V1 and V4 are for lower level details like colour and shapes
FFA
Preferentially responds to images of faces
Bottom of temporal lobe
On the fusiform gyrus
Parahippocampal place area (PPA)
Preferentially responds to images of places, houses, scenes
Bottom of temporal lobe
In parahippocampal cortex
Extrastriate Body Area (EBA)
Preferentially responds to pictures of bodies and body parts
It is further back but still on ventral pathway
Part of the extrastriate cortex
