Lecture 12 Object Recognition Flashcards
(23 cards)
What is the role of the inferior temporal (IT) cortex in vision?
It plays a key role in object recognition
it’s part of the ventral ‘What’ pathway, integrating form, colour, and depth.
What features are detected at lower levels of visual processing?
Bars and edges via LGN and V1.
Why is object recognition important for survival?
It helps in identifying food, avoiding danger, and navigating the environment.
What is Marr’s computational theory focused on?
Understanding object recognition via detecting edges, lines, and curves.
What brain area detects edges according to Marr’s model?
LGN (Lateral Geniculate Nucleus)
What brain area detects lines?
V1
What brain area detects curves and surface shapes?
V4
What kind of representation does V4 provide for objects?
Viewer-centered representation (template matching)
How does object recognition become viewpoint-independent?
Higher visual areas integrate features into a coherent, stable percept.
What is visual object agnosia?
Inability to recognise objects visually due to inferior temporal (IT) damage, despite intact vision and other senses.
What is prosopagnosia?
Face blindness — the inability to recognise familiar faces due to damage in the fusiform gyrus 梭状回.
What is a key feature of visual neurons in IT cortex?
Large receptive fields
o Robust response even when;
* object moves within the receptive field
* object changes in size
o poor response to simple stimuli such as spots or lines
What is the face inversion effect?
People struggle to recognise upside-down faces, showing faces are processed holistically.
What did Kobatake & Tanaka (1994) find in monkey IT neurons?
Face-selective neurons respond strongly to whole faces, but not to isolated features.
What is the FFA and what does it do?
Fusiform Face Area, specialised for face recognition (Kanwisher et al., 1997).
What did Gauthier suggest about the FFA?
FFA might be for expert object recognition, not just faces — we’re just face experts.
What are some modules of the ventral visual stream?
- V4 (form, colour),
- LO (Lateral Occipital, object recognition),
- FFA (faces),
- PPA (places),
- EBA (bodies).
What happens when the PPA is activated?
It responds to scenes/places — useful even in patients with object agnosia (e.g., Patient DF).
What kind of memory is scene recognition compared to?
Like face memory — involves detailed recognition, inversion effect, and distinct neural processing.
What is aphantasia?
The inability to form visual mental images, but other cognitive functions remain intact.
What distinguishes holistic from structural agnosia according to Farah (1990)?
Structural: based on parts/features; Holistic: based on whole configurations.
What evidence supports separate modules for object vs. scene memory?
Patient DF had object agnosia but could still recognise scenes, activating PPA normally.
How does hierarchical processing occur in object recognition?
Information flows from low-level features (edges) to complex object representations in IT cortex.
