Object And Face Recognition Flashcards
Challenges of object recognition
- different classes of things recognised
- from different angles-partly occluded
- occurs rapidly and mostly without error
Computers vs humans
Computers start with pixels and build up (bottom up) and people take info at multiple scales (top down)
Process of recognising an object
- See the basic features in the scene
- Perceive organisation in features
- perceive shape
- compare the shape percent to memory
- make a decision about whether the object is familiar
Template matching
Comparing the whole object to stored representations to find a match
Feature analysis
Break shapes down into critical features
- these critical features are recognised and assembled into objects and shapes that are compared to mental templates
Strengths of feature analysis
Reduces the number of representations the mind must store in order to process everything to be recognised
Marr and Nishiharas model of 3 levels
- Edge image -edges colours and blobs
- 2 1/2 D sketch - depth and orientation
- 3D model representation - relative positions
Better at explaining how computers are programmed than recognition in humans
Motion perception
We must be able to pick up changes in a stimulus over quite brief time intervals and recognise the same stimulus in different positions
Face recognition
Evidence that it is “special”
1) Developmental- newborns averaging 8 minutes old oriented towards upright face stimuli
2) Configural processing- see very specific spatial relationships between features
3) Mechanistic - special brain areas devoted to face processing - Fusiform Face Area (FFA)
Prosopagnosia =face blindness
- brain damage to face processing regions
- harder to do as involves making finer distinctions
Capgras Syndrome
A person holds delusions that friends and family have been replaced by an identical looking imposter
- damage of emotional pathway -people might “look” real to you but not “feel” real
Implications of face recognition.
- facial attractiveness related to more lenient court decisions for crimes where attractiveness is a factor e.g. Con artist
- “criminal” faces rated as less trustworthy and more dominant