Catégories : Tous - bias - recognition - processing - expertise

par Joerg Bauer Il y a 17 années

573

Face recognition (within category judgement)

The text explores the nuances of face recognition, contrasting it with object recognition and discussing the specialized processes involved. It highlights the inversion effect, noting how face recognition is disrupted more than object recognition when images are inverted.

Face recognition (within category judgement)

Face recognition (within category judgement)

Bruce & Youngs ("Square" model)

_Burton & Bruce IAC
connectionist model
Two Routes?

Familiar / Unfamiliar=different or just harder

Malone: Doulbe dissistiation with 2 Patients (famous statesmen)

Unfamiliar

Structural

Expression

Speech

Directed visual processing (difference to Familiar to check specific facial detail)

Familiar

Structural: its face

Face Recognition (i know him)

Identity nodes (its a singer)

Name generation (its mick jagger))

How?
Evidence

3 Stages: Bruce Rt's

b) x is a politician faster than naming

a) Familiarty is faster than sematnic (occupation)

Object / Face diference

Double dissitiation

Object recognition

Agnosia: "Outpouchings" (Oliver Sacks)

Face recogntion

Propagnosia: some dont recognize faces

Naming=different

Young, Hay Allice: Diary Study: we can't name unless semtic info

5.Cognintve System : Evaluate and feed back on what we see
4. Directed visual Processing: Other semantic information (singers are attractive- check out his eyes
3B. Facial Speech analyis
3A Facial Expression (emotional) analysis
2 C. Naming (its Mick Jagger)
2 B. Person identity nodes PIN (a musician)
2A. Face Recogntion units (FRU: i know him)
1.Structural encoding (its a face?)

Viewpint indendent (3D) descriptions

Viewpoint (2D) dependent

Issues with explanation of face recognition

Neurological issues
Capgras delusion both for face and objects

Damaged ventral

Prosppagnosia

Bauers Covert Recognition

GSR response

IAC expalins this as weekend FRU to PIN connections

PIN to SIU ok (Therapy)

Points to damaged Ventral (what)

Youngs ex-serivcemen

Familiar / Unfamilar reogntion may be affected

Facial Expression Analysis may still work

Two Routes: Dorsal / Ventral

Objects and Faces in different areas
fMRI
Double dissotiations
More holisitc
Grotesque

Inverted impaired more 2nd order features: Epert tuning lost

1st order: eyes, nose, mouth

But dog experts have more problems with inversed dogs
Yins Inversion effect is more destructive for faces comp to objets
Difference familiar / unfamiliar faces
Matching rate only 80%
We need to account for facial emotions and facial speech
Its a within categ. comp - with same cat. objects
Facees all look similar, objects dont
Face Experts: innate bias toward faces:
Same three stages as object recognition: Percept.recog , sematic knowlege, naming
Faster Rts of famil vs. semantic
Evidence Youngs diary sutdes

Same processing for face vs. object recognition

Quoalitatively Different: Yin: Inversion effect not for objects ?
Holistic/ configural processing
Diamond & Carey: Dog experts vs. non Experts worse at inversed dogs

Second order: Differences in those relations

Evidence: Searcy & Bartlett: Inversed spatial relationships were less grotesque

First order: eyes above nose

Innate/ but Kick Start Johnson & Morton: Newborns show attentional bias to faces
Location specific: Evidence for a "Granmother" cell

Is it different from Object Recog?

Is it just different because of within/ between category classification (unfair comparison)
Neuropsychological Evidence
Different Locations (fMri) in the brain from object recog
Double Dissotiation

Agnosia

Damage at earlier stages visual procc: A brush: "A wooden stick and a black object" (Humprheys)

Inability to rec objects ("outpouchings")= stop at percept (2,5 D or 3D) but no recogn)

Prosopagnosia

Inahbility to recognize familiar faces

Capgras Delusion

Burton & Bruce IAC (Interactive activation and competition)

Name Recog
NRU's (Name Recog Untits)
WRUs (general word recgo units)
Visual Recog
Say: its M. J.
PINS (Personal Id nodes) One node matches a person
SIU's (It's a singer and his name is Michael J.)
FRU's (Face Rec Units)= sensory id

Bruce and Youngs General Model of Face Recog

Facial Speech Analysis (Lip reading)
McGurkEffect (ba, ba, ga)
Facial Expression Analysis
Face Recognition
Name Gerneration
Person Identiy Nodes (Semantic knowledge)
Face Recognition units (Familarity)
Yin: Immediate recognition is good - probably image recognitino, not face
Cognitive System : Overrule decisions, if unsure

Face perception Organisation (Stages)

We have expertise through attentional bias from birth
Highly difficult task: High leval of detail
Bruce and Johnstons Speed Measuring: Recogntion happens in Stages
1. Perceptual classification, 2. Sematic classification, 3. Naming
Starts where perception ends=structural classification

Familiar / unfamiliar face rec. is different

IAC explains this: FRU to SIU-Not enough activation for covert/ enogh for overt
Concious (overt) vs. covert recog. points to seperate neural pathways
Overt Ventral
Covert Dorsal
Neuropsychological Evidence for indepent Modules by: Young: ex-servicemen
Unfamiliar Faces
Generally Worse

Influenced by pose, lighting and viewpoint

Bruce: Video Line Up 80% correct

Kemp: Credit card image matching 36%correct

Familiar Faces (Paul McCartney)
Change independent
Time independent
Viewpoint independent