Visual Perception and Cognition Part 2: The Architecture of Vision Flashcards
Perception is the study:
Of the processes of the organization of the external world as encoded by our senses
What is the main issue about perception?
The big issue is how we encode (recognize, categorize, act upon..) the world
What do we mean by “perception is like a reflex”:
computational mechanisms that act like they are “smart” but which are “dumb”
Perception - Why “smart”?
– Because it knows a lot about how to encode properties of the world; how to “interpret” the world
– It has rules and representations
Perception why “dumb”?
– Because in large part it operates without relying on higher cognitive mechanisms; it “does its job” automatically, like a reflex…
– It doesn’t “think” about what it does
What are two examples on how the perception system is “smart”?
(1) Assigning 3-D structure to 2-D “objects” based on line arrangements and vertices
(2) It can extract shape from shading and motion
Face perception (convex-like form) is
also automatic
Shape-from-shading is
a perceptual phenomenon where the brain deduces the three-dimensional shape of an object based on the patterns of light and shadow it casts. When a surface is illuminated, the resulting shadow can provide important information about its curvature. For example, a convex object will show a pattern of light and shadow that the brain interprets as consistent with an outward curve. This process occurs automatically, often without conscious effort, allowing for rapid and accurate object recognition.
Face perception is
a specialized domain where the brain is particularly adept at recognizing convex-like forms. Faces, for instance, are typically perceived as convex, with features like the nose, cheeks, and chin projecting outward. The brain has evolved to automatically process these convex cues, enabling fast and efficient recognition of human faces.
The visual system relies on :
regularities of the world (probably hard-wired)
Light shines from
“above” producing two different 3-D representations (convex and concave)
.
Many properties of the world seem to be:
Encoded, wired in
The principles (/rules) encoded in the visual system
Make its operation automatic
Some of the principles/rules operate
in actual object and face recognition
Why do we say that perception is “dumb”?
If the visual system doesn’t “think” about what it does (i.e., if it applies its principles - rules - without relying on higher knowledge systems), it can make mistakes
What is an example of perception being “dumb”:
Even when you know the properties of a given stimulus, the visual systems
persists with its own “interpretation”
What are some examples of perception being “dumb”
It makes assumptions on:
Relative size
Orientation
Trajectory/movement
Shape
Convexity
texture of objects and scenes
The persistence of visual illusions:
Visual illusions keep re-occurring despite our knowledge that they are illusions!
what is the muller lyer illusion
The Müller-Lyer illusion is a well-known optical illusion involving two lines of equal length, each with different arrow-like endings. In the illusion, one line has outward-facing “fins” or arrows at each end (like this: <—>), while the other line has inward-facing fins (like this: >—<). Although the lines are the same length, people typically perceive the line with outward-facing fins as being longer than the line with inward-facing fins.
The inverse-optics problem:
A single image (a square) can be produced by an infinity of real-world objects
The inverse-optics problem: a single image (a square) can be produced by an infinity of real-world objects
They all produce the same retinotopic representation
The world as it is:
Three-dimensional objects and scenes
Low-level representation:
Arrays of light intensities
