Face Matching Flashcards
4 categories to this revision, what are they?
The Bad
The Good
Training
AI
What four factors make us worse at recognising unfamiliar faces?
- Lighting
- Hair style
- Camera angle
- Expression
What did Noyes & Jenkins (2019) find and conclude about disguised faces?
Disguised familiar faces accuracy dropped in the same familiar face category
Unfamiliar face recognition dropped in ALL three categories: same, different similar, and different random, when non-disguised face were compared to disguised faces.
Concluded: we are bad at recognising faces when disguised
What is the Other-Ethnicity effect (Zhoe et al., 2021) and how does that impact our ability to recognise faces?
The Other-Ethnicity effect is us being bad at recognising faces of other ethnicities compared to our own. However, growing up in a multicultural environment mitigates against this effect.
What did Zhou & Mondloch (2016) conclude from their Ethnicity Sorting task?
Chinse people and White people tested on the Other-Ethnicity effect.
Found that we are bad at recognising unfamiliar faces that were of a different ethnicity. Made more piles of faces for unfamiliar faces than familiar faces
What is the the second, “Other ____ effect”? (Kuefner et al., 2008)
Other “Age” effect
Older better at recognising older faces, young better at recognising young faces
For recognising faces, what is the overall conclusion based on all the evidence?
We are bad at recognising faces when they are:
Disguised (Noyes & Jenkins, 2019)
Other Ethnicity effect (Zhou & Mondloch, 2016)
Other Age effect (Kuefner, 2008)
When familiar faces are disguised are we good or bad at recognising them?
We are good at recognising familiar faces when disguised
What did White et al., (2014) study conclude about Passport Officers in their ability to recognise faces?
They were not good. Experience had no impact on their ability to recognise similar faces.
What did White et al., (2015) conclude about Forensic Examiners and their ability to recognise faces?
Tested 3 groups; undergraduates, controls and examiners.
2 images of same face, 2 images of different faces.
Examiners performed better than the other two groups in 2 out of the 3 tasks.
What did Yates et al., (2023) find in their study about Forensic Examiners when the faces were disguised?
Examiners were good at identifying unfamiliar disguised faces
In Fysh & Bindemann’s (2022) training study what did they test?
Training individuals to recognise highly identifiable marks
In Fysh & Bindemann (2022) study what were their findings on training individuals to recognise highly identifiable marks (moles, birthmarks)?
2 same images, 2 mismatch images.
3 trial types, control, mole congruent and mole incongruent.
When moles were congruent performance was greater than the other two trials, in recognising same and mismatch faces.
In Fysh & Bindemann (2022) follow up study (to the highly identifiable marks one) what did they do and what did they find?
They instructed participants to use the highly identifiable marks, moles, to identify faces.
They found performance in identifying faces was even greater than the first study. Comparing the two studies there was an 18-21% increase when given instructions.
What did Towler et al., (2021) conclude with pre training?
Participants split into 3 groups, control, non-diagnostic, and diagnostic.
2 same images, 2 mismatch unfamiliar faces
No benefit of pre training in control and non-diagnostic training, in-fact they performed worse post-training. However, the diagnostic group was significantly better at matching faces, post-training.
