5: Perceiving Objects Flashcards
What are two key difference between machine “vision” and human vision?
Machines cannot identify objects (e.g., bush vs. rock).
Machines navigate in 2 dimensions well, humans in 3 dimensions.
What is the inverse projection problem? Why is object perception difficult?
Stimulus falling on the retina is ambiguous.
Projection of 3D world onto the 2D retina, then reconstructed into a 3D perception.
What does the inverse projection problem show?
Different objects can cast the same image on the retina. Same object casts different images on the retina when viewed from different angles.
The ability to recognize a single object seen from different viewpoints is called?
Viewpoint invariance.
Perceptual organization involves what two components?
Grouping: elements are put together into objects.
Segregation: separating one area/object from another.
Wilhelm Wundt founded what? Explain how perceptions are created under this theory and how objects are perceived.
Structuralism.
Perceptions created by combining individual sensations.
Objects perceived by grouping specific sensations, and segregating them from other things.
“The whole is the sum of its parts” refers to what?
Structuralism.
What are the Gestalt Laws of Perceptual Organization?
Pragnanz: law of good figure / law of simplicity. Every stimulus perceived so structure is simple as possible.
Law of similarity: similar things appear grouped together.
Law of good continuation: lines and patterns perceived as continuing in time and space. Straight or smoothly curving lines are seen as belonging together.
Proximity of Nearness: things that are near to each other appear to be grouped together.
Law of common fate: objects moving in the same direction tend to be grouped together.
Name three “rules” of Gestalt perception psychology.
Common Region: elements within the same region appear grouped together.
Uniform Connectedness: connected regions are perceived as a single unit.
Synchrony: events that occur at the same time are perceived as belonging together.
Describe figure-ground segregation.
Figure: a separate object, that stands out from the background.
Ground: the background on which the object sits.
What are two properties of the “figure” in figure-ground segregation?
Figure perceived as more “thing-like,” more memorable than the ground.
Figure is usually seen as being in front of the ground.
In figure-ground segregation, the “ground” is perceived as what?
Perceived as “unformed,” and extends behind the figure.
What is border ownership?
The border that separates the figure from the ground is perceived as belonging to the figure.
What four things tend to be perceived as figures?
Symmetrical areas, convex things, small things, things on bottom of scene.
Regarding Gestalt laws and “meaningfulness,” what happens during prolonged viewing?
Familiar objects tend to be grouped and segregated from the ground.
A scene is acted ____, while an object is acted _____.
Within; upon.
How good are people at perceiving scenes?
90-100% accuracy in detecting target image when array of images are presented for ¼ of a second.
We make use of environmental regularities to obtain global image features. What are the five global image features?
Degree of Naturalness: natural vs. man-made
Degree of Openness: open space vs. cluttered
Degree of Roughness: # of small elements/clutter
Degree of Expansion: visible horizon
Color: characteristic colors
Regularities based on the meaning of the scene are known as what?
Semantic regularities.
What is the theory of unconscious inference?
Some perceptions are the result of unconscious assumptions we make about the environment.
What is the likelihood principle?
We will perceive an object that is most likely the cause of our sensory stimulation.
Why are people better at object perception than computers?
Application of knowledge to perception.
Regarding face perception, what are the roles for:
- V1
- FFA
- Amygdala
- Superior temporal sulcus
- Frontal lobe
V1: start of cortical processing.
FFA: specialized face identity module.
Amygdala: emotional aspects of the face.
Superior temporal sulcus: gaze direction, speech movements.
Frontal lobe: attractiveness.
Describe training and testing decoders during pattern recognition tasks.
Training: what voxels become active, knowing what picture the subject is viewing.
Testing: measure the active voxels and guess which stimulus is present.
How did Jack Gallant’s lab develop a more complex brain reader?
Modeling the spiking activity of neurons that may lead to the observed BOLD response. Seeded model with images from YouTube. Implemented a semantic model.
