cognitive psychology - face recognition

eyewitness identification

• 1970’s research was mainly interested in recognition of unfamiliar faces due to several cases of misidentification

• eyewitness misidentification contributes to an overwhelming majority of wrongful convictions that have been overturned by post-conviction DNA testing - 60% of their clients were wrongfully convicted based on eyewitness misidentification

when is face recognition used

• passport control

• ID for age verification

• police monitoring crowds

• airport face scans

• video surveillance

• car and mobile face recognition systems

• prosopagnosia - bates and bennetts 2014

• autistic spectrum disorder - stantic et al 2021

what did bahrick find - 1975

• remember names and faces of school friends over long periods of time - hardly any forgetting over 35 years

what did bahrick find - 1984

• teachers recognition of students taught 3-5 times for 10 weeks

• recognition of recently taught - 69% dropped with intervening years

• 8 years later 26% of students recognised

recognising unfamiliar faces

• recognition of faces seen once or twice not accurate

• unfamiliar face recognition is influenced by viewpoint, expression and environmental context

what was the kemp et al study

• supermarket cashiers challenged around 10% of people presenting a valid card and accepted 64% of the invalid cards if there was some similarity between the card presenter and the person pictured - 1997

what was the white et al study

• passport officers in sydney accepted fraudulent photos 14% of the time - 2014

what is recognising familiar faces influenced by

• changes in appearance

• angle

• lighting

• age

• hairstyle

what is recognising unfamiliar faces influenced by

• viewpoint

• expression

• environmental context

what was the yin study

• immediate testing - participants do well, 93% but if photo shown at test is a different angle or expression the recognition rates drop - 1969

what were the results of the study by young et al

• person misidentified - 314

  • someone unfamiliar misidentified as someone familiar

• person unrecognised - 114

  • someone familiar thought to be unfamiliar

• person seemed familiar only - 223

  • recognise someone as familiar but no other info

• difficulty retrieving full detail - 190

  • only some semantic info retrieved, not their name

• decision problems - 35

  • think you recognise but decide it cant be them

bruce and young model of face recognition - 1986

supporting for model of face recognition

• sequential nature of recognition system supported

  • hay et al 1991

  • johnston and bruce 1990

refuting for model of face recognition

• recognition of unfamiliar faces not well explained

• people can match faces and their names without accessing any semantic information - e.g occupation

• de haaan, young & newcombe 1991

valentines face space model 1991

what does valentines face space model suggest

• explains why some faces are easier to recognise

• distinctive faces more accurately and quickly recognised than faces seen as ‘typical’

• representations of distinctive faces are stored in more sparse areas of face space where they are not confused with other faces

what are the factors affecting facial recognition (7)

• facial motion

• distinctiveness

• race

• age

• retention interval

• forensic factors

• weapon focus effect

what is facial motion

• movements of a face can facilitate more accurate and faster recognition - xiao et al 2014

what is the supplemental information hypothesis for facial motion

• characteristic facial motions exhibited by a face are stored within the facial representation of that individual - O’toole, Roark, Abdi 2002

what is the representation enhancement hypothesis for facial motion

• motion facilitates the perception of 3d structure of the face - O’toole, Roark, Abdi

negatives of facial motion

• inconsistencies have been found using unfamiliar face stimuli - Lander & Bruce 2003, Pike et al 1997, Christie & Bruce 1998

what is said about distinctiveness of faces

• faces rated as distinctive are recognised more accurately than faces rated as typical

  • Cohen & Carr 1975, Light, Kayra-stewart & Hollander 1979, Newell, Chiroro & Valentine 1999

what did butcher & lander find about distinctiveness 2017

• found that the more distinctively a face moves the easier it is to recognise, and the more the face benefits from being seen moving

what is said about suspects and distinctiveness

• suspects who stand out and are distinctive are prone to be selected

  • Wells, Rydell & Seelau 1993, Wixted & Mickes 2014

• in the US, England and Wales, police replicate the distinctive feature across lineup members or conceal it

  • Police & Criminal Evidence Act 1984 Code D 2011, Technical working group for eyewitness evidence 1999

what is the other race effect

• more accurate recognition of faces of the same race as the observer than faces of another race

who came up with the other race effect

• Malpass & Kravitz 1969

model for race effect on facial recognition

what is said about children recognising faces

• children have poorer recognition, smaller inversion effect, recognise typical and distinctive faces equally

  • Valentine, Lewis & Hills 2016

• children can be as good as adults, but more easily disrupted by changes to testing parameters

  • Want, Pascalis, Coleman & Blades 2003

what is said about age affecting facial recognition

• older adults have poorer face recognition

  • Searcy et al 2001

• Evidence of an own-age bias in children and older adults

  • Anastasi & Rhodes 2005

• Stimuli age - close to change performance when recognising the same person across a 10-30 year age gap

  • Sexton et al 2024

what is the retention interval

• the amount of time (delay) between initially seeing the face and being to recognise or remember it

retention curve model Ebbinghaus 1885

positives of face retention interval for facial recognition

• examined face recognition accuracy using retention intervals of 4 weeks or more

  • Courtois & Mueller 1981, Sauer et al 2010, Shepherd & Ellis 1973, Shepherd et al 1982, Yarmey 1979

• A statistically reliable association between longer retention intervals and decreased face recognition memory has been found, during both face recognition memory tasks and eyewitness ID studies

  • meta-analysis by Deffenbacher et al 2008

what did Kramer 2021 find about retention intervals

• the majority of forgetting took place within the first 24 hours, with no significant decrease after that timepoint

  • no delay, 6 hours, 12 hours, 1 day or 7 days

what did Sauer 2010 find about retention intervals

• longer delays of several weeks also lead to greater overconfidence on lineups than immediate lineups

what are estimator variables for forensic factors

• intrinsic to person and situation they were in

what are examples of estimator variables

• age, race, intelligence, gender, personality, confidence, stress

• illumination at the scene, presence of a weapon, exposure duration and delay

what are system variables for forensic factors

• procedures used to obtain an accurate identification

• often under control of legal system and could be altered

what are examples of system variables

• prior exposure to mug shots

• composition of line up

• moving or static photographic line ups

• number of fillers

what is said about the weapon focus effect

• eyewitnesses observing a crime in which a perpetrator carries a weapon are less accurate in describing or identifying the suspect in a lineup

positives of weapon focus effect

• 42/63 experts said the effect is generally or very reliable

• 87% said its reliable enough to testify in court - Kassin et al 2001

negatives of weapon focus effect

• many have found a WFE but not all

  • Cutler et al 1987, Hope & Wright 2007, Johnson & Scott 1976, Shaw & Skolnick 1999

• meta-analysis found a WFE effect on perpetrator/target descriptions but not perpetrator identifications

  • Kocab et al 2016