Object Recognition Flashcards
What are three processes of Marr’s computational model of visual processing.
- Edge detection
- Orientation coding (simple cells)
- Curvature and shape (V4 responsible)
What is the “What” pathway also known as?
The ventral stream.
Name the modual ordering of the visual streams model? Where and what is the Inferior Temporal (IT) cortex?
- LGN, V1, V2, V4, IT
- IT functions as high-level structual representation.
What is one difference between V4 and IT?
IT is not viewpoint dependent. i.e involves larger networks like learning and memory
What is visual object agnosia? How can it occur?
- Loss of an ability to recognise familar objects.
- Damage to the inferior part of the temporal cortex (IT).
Specifically visual forgetting. Not related to other senses
What did Farah (1990) argue for regarding visual agnosia?
Argued for two types:
- Structual mechanisms (based on identifying features)
- Holistic mechanisms (based on identifying configurations)
Consistent with seperate modules for faces and other familiar objects.
What does the face inversion effect imply?
Faces upside down are harder to recognise.
What did Desimone (1984) show?
(Faces)
That there were neurons in the inferior temporal cortex that responded specifically to faces.
masking specific facial features did not completely zero the response.
What did Kobatake & Tanaka (1994) show in primate IT ?
// facial features
- Showed cells that are most responsive to intact faces.
- Cells not responsive to facial features presented in isolation.
What did Tsao (2006) show?
// faces
- Use fMRI to record 500 single cells in ‘face’ area.
- 97% of visually responsive cells responded more to faces than a range of control groups.
Kanwisher (1997) work is linked with what region of the brain?
Fusiform Face Area (FFA).
What is the alternative hyp’ of the FFA?
Gauthier.
That the FFA actually relates to expertise.
What did Bell (2009) provide evidence for?
// faces and body parts
That faces and body parts activated adjacent regions.
What can damage to the FFA induce?
work by Uttner (2002)
Face blindness (prosopagnosia).
What did Standings (1973) show?
// natural scenes
- people shown 10,000 scenes for a few seconds each
- Could later recall with 83% accuracy
Basic yes or no recall
What did Brady’s (2008) repeat detection task show?
// objects
- That our encoding of object information into memory is very good.
- Three conditions each of which tested pairs of objects that were more or less similar. Strong retention for all of them.
Konkle (2010) and visual scenes showed?
- Increasing category exemplars creates a higher completion for memory resources.
- Decrease of memory performance.
Memory still well above chance with higher exemplars
Same results for higher exemplars in object categories
When performing memory tasks, what is seen when we increase the number of exemplars in any given category?
- A decrease in memory performance due to a competition of memory resources.
- This is found for object and natural scene recall.