Visuelle Wahrnehmung
Quiz Fragen zu der Vorlesung
Quiz Fragen zu der Vorlesung
Fichier Détails
Cartes-fiches | 151 |
---|---|
Langue | English |
Catégorie | Informatique |
Niveau | Université |
Crée / Actualisé | 01.11.2017 / 27.01.2018 |
Lien de web |
https://card2brain.ch/box/20171101_visuelle_wahrnehmung
|
Intégrer |
<iframe src="https://card2brain.ch/box/20171101_visuelle_wahrnehmung/embed" width="780" height="150" scrolling="no" frameborder="0"></iframe>
|
7: What is unique about face perception and how is it different than object perception?
Faces are different than other objects because all faces have the same parts in the same relationships with one another (e.g., eyes above nose, which is above the mouth). Therefore, fine metric details of faces are important in recognition, and it seems the visual system represents faces holistically in terms of these fine metric details, whereas it does not in the case of objects. Further evidence that the visual system treats faces and objects differently is the double dissociation between face and object recognition regions of the brain. Some patients with brain damage develop object agnosia and cannot recognize objects but can still recognize faces. Other patients develop prosopagnosia and thus cannot recognize faces but can recognize other objects. Finally, inverted faces are much harder for us to recognize than inverted objects, suggesting that faces are processed differently than objects.
7: What kinds of processes happen in middle vision?
Middle vision refers to a set of processes that combine features detected in early vision (such as edges and contours) into objects. Middle vision utilizes rules and principles for combining elements into perceptual groups, many of which were discovered by psychologists from the Gestalt tradition. Some important steps in middle vision include finding edges of objects, dealing with occlusion, texture segmentation and grouping, and determining figure/ground assignments.
8: Evidence indivates that strucures in --- cortex are especially important in end-stage object recognition processes
inferotemporal
Tipically ...
which is a entry-level category term?
Bird
8: --- is a failure to recognize objects in spite of the ability to see them
Agnosia
8: What are object representations made of, according to the recognition-by-components (RBC) model of object recognition?
Geon structural descriptions
8: What are object representations made of, according to view-based theories of object recognition?
Image templates
8: Many researcher is vision science believe that object recognition is one of the most important functions of the human visual system. Thus it is perhaps not surprising that there exists a large body of research on object recognition.
One dominant approach - usally referred to as --- models - postulates a "visual alphabet" made from 3D geometric primitives.
structural descrption
8: Many researcher is vision science believe that object recognition is one of the most important functions of the human visual system. Thus it is perhaps not surprising that there exists a large body of research on object recognition.
One dominant approach - usally referred to as structural description models - postulates a "visual alphabet" made from 3D geometric primitives.
The most prominent theors of this kind is called ---, published in --- by ---.
recognition-by-components
1987
Biederman
8: Many researcher is vision science believe that object recognition is one of the most important functions of the human visual system. Thus it is perhaps not surprising that there exists a large body of research on object recognition.
One dominant approach - usally referred to as structural description models - postulates a "visual alphabet" made from 3D geometric primitives.
An oppoing theory - usally referred to as --- models - instead believes the human visual system recognizes objects by storing and matchin "2D-images" or "snap-shots" of objects and inter- and extrapolates between them if required to recognize an object from a novel viewpoint.
view-based
8: Many researcher is vision science believe that object recognition is one of the most important functions of the human visual system. Thus it is perhaps not surprising that there exists a large body of research on object recognition.
An opponing theory - usally referred to as view-based models - instead believes the human visual system recognizes objects by storing and matchin "2D-images" or "snap-shots" of objects and inter- and extrapolates between them if required to recognize an object from a novel viewpoint.
One of the most well-known proponents of this theory is ---. --- both theories are --- connected to the findings, theories and models of early spatial vision.
Tarr
Unfortunately
not
8: Many researcher is vision science believe that object recognition is one of the most important functions of the human visual system. Thus it is perhaps not surprising that there exists a large body of research on object recognition.
Very recently DiCarlo, Cox and colleagues have argued for a computational neural netword approach to explain object recognition. This can be seen as a neuroscience-inspired computational instantiation of the --- approach to object recognition.
view-based
8: Some researchers believe that a major problem with structural decription theories of object recognition is that
structural descritpion theories predict that object recognition should usally be viewpoint invariant, but in fact recognition has been shown to viewpoint dependent.
8: A major problem with naive template theories of object recognition is
that we cannot possibly stor enough templates in memory to match every object we might encounter
8: One central property of the generalized-cone components is that they have so-called --- properties, because these properties would --- be produced by --- alignments of viewpoint and object features.
non-accidental
rarely
accidental
8: Thus RBC theory claims that certain properties of the --- image are taken by rhe visual system as strong evidence that the edges in the --- world contain the same properties.
2D
3D
8: The essential non-accidental properties of the generalized-cone components are:
8: Prosopagnosia is a neuropsychological disorder in whch the patiant
cannot identify faces, but can recognize other types of objects
8: A study of cells in IT cortex of ahuman patien showed that they responded to very specific stimuli, such as ---.
celebrities
8: Which is a loosely defined stage of visual processing that comes after basic features have been extracted from the image, and before object recognition and scene understanding?
Mid-level-vision
8: Which is a subordinate level category term?
Limousine
8: Which is a superordinate level category term?
Vehicle
8: Tarr and his colleagues found that the amount of time needed to recognize novel objects is at least partially determined by
the degree to which the object is rotated from ist studied view
8: Viewpoint invariance refers to the idea that
objects schould be just as easy to recognize from any viewpoint
8: Yasmins and colleagues from the DiCarlo lab at MIT published an artivle in 2014 in which the presented their HMO model, standing for ---
hierarchical modular optimization
8: Yasmins and colleagues from the DiCarlo lab at MIT published an artivle in 2014 in which the presented their HMO model, standing for hierarchical modular optimization.
The HMO model belongs in the larger class of --- models, standing for --- model.
DNN
deep neural network
8: Yasmins and colleagues from the DiCarlo lab at MIT published an artivle in 2014 in which the presented their HMO model, standing for hierarchical modular optimization.
The HMO model's essential architectural characteristic is its --- : There are, for example, many --- connections and different parameter ---.
heteroogeneity
bypass
settings even at the same level
8: Yasmins and colleagues from the DiCarlo lab at MIT published an artivle in 2014 in which the presented their HMO model, standing for hierarchical modular optimization.
---, the basic operations performed locally are --- troughout the network.
However
the same
8: Yasmins and colleagues from the DiCarlo lab at MIT published an artivle in 2014 in which the presented their HMO model, standing for hierarchical modular optimization.
In the Yamins et al. (2014) article they report a large-scale modelling effort, evaluating around ---. Yamins et al. compared their models both to the response of cells in IT cortex (roughly N = --- cells) as well as on how well the models categorized a set of images (roughly N = --- images). One central finding was that models optimized for --- were also superior at --- .
5000DNN architectures
300 oder 100
6.000
categorization performance
explaining variance in IT
8: Yasmins and colleagues from the DiCarlo lab at MIT published an artivle in 2014 in which the presented their HMO model, standing for hierarchical modular optimization.
In order to obtain a categorization performance from the HMO model, a --- decoder was --- the activity of units at the --- level(s) of the HMO network. Using such a procedure, the HMO model's performance was --- than that of --- models of object recognition on the difficult --- variation task.
linear
trained on
highest
better
both computer vision and neuronally inspired
high