Face Recognition Flashcards Preview

Cogniton In Clinical Contexts > Face Recognition > Flashcards

Flashcards in Face Recognition Deck (19):

First order relations

All faces share a basic recognition, features which are in the same place


Second order relations

Where the features actually are on the face (spacing etc)


Early model of face recognition - Bruce and young

Modular model -different functions are processed independently. Structures in the brain support different functions
There are distinct pathways for recognising familiar faces vs doing other things (expressions)


Parallel processing

Dealing with expression, speech etc


Serial processing

X happens before Y can
E.g. Working out someone's face, then name


Route to recognising faces

Name generation units



Face recognition units - shared descriptions of previously encountered faces



Personal identity nodes - specific info relating to that person (likes etc)


Name generation units

Taking input from pins, to generate name of individuals


Early evidence for Bruce and young

Memory loss diary study (young hay and Ellis) - ppl show different types of errors which suggests breakdowns at different points in facial recognition process - suggests a serial process


Other routes to recognising faces

Expression analysis - computing facial emotions such as happy, fearful, happens in parallel
Facial speech analysis - lip reading, happens in parallel
Directed visual processing - when looking for particular information from a face


Repetition priming

If you've seen something/heard it, the next time you're quicker to process it
Found for familiarity decision but not for gender or expression decisions


Brain support for Bruce and young

Particular deficits in the brain, of you have one ability entact but not one it shows there are seperate modules
Some people can identify people but not their emotions
Different parts of brain are active when doing different things


Semantic priming - key limitation of Bruce and young

A face responds faster if it follows a closely related face
Bruce and young don't account for this


Interactive activation and competition model

McClelland proposed parallel distributed networks a that have interactive activation and completion build it
Info stored in concepts and categories
Semantic info is pooled, knowledge is represented in pools - nodes is all in one pool
Relationships between knowledge are represented in the connections between pools


What are connections within pools

Mutually inhibitory - one thing inhibits another eg if one name is activated, harder to remember another


What are connections between pools

Mutually facilitatory - a name will lead to activation of another pool
There is evidence things are grouped e.g. Vegetable names all in one group


Face selective neurons

Face neurons in superior temporal fulcas and inferior temporal cortex - single neurons respond to specific features


Grandmother cells

Existence of a single neuron to detect and represent a single object such as your mother

Patients with epilepsy had electrodes. Responses of neurons to images showed an Neuron responding to pictures of Anniston but different Neuron for pitt