Recognizing Objects Flashcards
What is bottom-up processing?
Directly shaped by stimulus, data-driven
What is top-down processing?
Processes shaped by knowledge, concept-driven
What are visual search tasks?
Tasks in which participants examine a display and judge whether a particular target is present in the display or not
Why are visual search tasks useful?
Support the claim that recognition begins with identification of features
What is a tachistoscope?
A device designed to present stimuli for precisely controlled
amounts of time
What is a mask?
A random pattern of lines, curves, or letters that is meant to interrupt any continued processing of the previously-presented stimulus
What is priming?
Preparing a participant for exposure to a stimulus
What is repetition priming?
Using an initial exposure to a stimulus to prepare the participant for another exposure to the same stimulus
What is the word-superiority effect (WSE)?
Letters are more easily recognized in the context of a word than in isolation
What is well-formedness?
How well a letter sequence conforms to typical English spelling patterns (produces similar effect as the WSE)
What are the common recognition errors?
Less common patterns (i.e. “TPUM”) are more likely to be misread as more common patterns (i.e. “DRUM”, “TRUM”), but the reverse effect doesn’t occur
Misspelled words, partial words, and non-words are adjusted to fit typical spelling patterns
What are feature nets?
A network of feature detectors, with the scale of objects increasing when moving upwards through the feature net
What is a detector’s activation level?
How energized the detector is/the status of the detector at the moment
What is a detector’s response threshold?
The point at which the detector will fire/send its signal to the other detectors to which it is connected
What are bigram detectors?
Detectors of letter pairs, triggered by lower-level detectors and send their output to higher-level detectors
How are ambiguous inputs interpreted?
The detector that is more primed will respond (i.e. “TAE” vs. “THE” - detector for “THE” has been primed significantly more than “TAE”, therefore “THE” will be recognized)
How do recognition errors occur?
Similar to the recognition of ambiguous inputs (i.e. if “CQRN” is presented, “CORN” will be recognized due to more priming)
What does it mean for knowledge to be “locally represented”?
Able to look at a single bigram and determine the frequency of occurrence
What is a distributed representation of knowledge?
Activation levels have to be determined by viewing the relationship between priming levels (i.e. “CO” priming levels need to be compared to “CF” priming levels)
How do we maximize efficiency and accuracy?
English has many redundancies, therefore not every letter of a word needs to be read in order to be able to correctly recognize the word (i.e. “contxt”)
What is the McClelland and Rumelhart Model?
A different representation of a feature network that involves both excitatory and inhibitory connections
What are excitatory connections? (McClelland and Rumelhart)
Connections that allow one detector to activate its neighbors (i.e. activation of “T” detector allows activation of “TRIP” detector)
What are inhibitory connections? (McClelland and Rumelhart)
Connections that allow one detector to inhibit its neighbors (i,e, activation of “G” detector inhibits activation of “TRIP” detector)
What is the Recognition by Components (RBC) model?
A feature network to explain the recognition of objects, containing an immediate level of detectors sensitive to geons
What are geons (geometric ions)?
The building blocks of all objects we recognize (the alphabet from which all objects are constructed), assembled into “geon assemblies” that activate the object model
What does it mean for object recognition to be viewpoint-dependent?
Only a certain number of viewpoints of a particular object are remembered, and those that are encountered more frequently create faster recognition than those less encountered
How do neurons in the inferotemporal cortex support the idea that object recognition is viewpoint-dependent?
Particular neurons in this area fire preferentially when a specific type of object is present in the visual field, and many of these neurons are view-tuned
What is prosopagnosia?
People cannot recognize faces of loved ones, famous people, or even themselves
What is the inversion effect?
The recognition of inverted objects is significantly less accurate than recognition of them in their typical orientation (this effect is particularly present in the recognition of faces)
What is holistic perception?
Perception that depends on the overall configuration of the object rather than an inventory of its parts (how faces are likely perceived)
