![\"Dr](\"https:\/\/www.open.ac.uk\/blogs\/news\/wp-content\/uploads\/2021\/11\/Catriona_crop_two.png\")
“Matching faces can be difficult and error prone,” says Dr Catriona Havard<\/p><\/div>\n\n\n\n
Unfamiliar face matching, that is, deciding if two different images of a face are the same person or different, is important for a variety of identification tasks, such as at passport control, or when buying alcohol or other restricted items needing photographic identification.<\/p>\n\n\n\n
Dr Catriona Havard, senior lecturer in psychology had her paper \u201cThe Importance of Internal and External Features in Matching Own and Other Race Faces\u201d, <\/a>recently published in the journal Perception. She said:<\/p>\n\n\n\n \u201cAlthough matching two faces may seem like a straightforward task, a wealth of psychological research has shown that, even from high-quality images, it can be difficult and error prone. And these are tasks which are often done by humans, in ID parades for example or identifying someone at a passport control.<\/em><\/p>\n\n\n\n \u201cIt\u2019s a bit of a myth that we\u2019re really good at matching faces, when in fact this shows we\u2019re not that good at all.”<\/em><\/p>\n<\/blockquote>\n\n\n\n In the study 388 participants took part, half in the UK and the other half in China, and all were split under gender lines.\u00a0<\/p>\n\n\n\n A collection of images was presented to the participants. Half the images were Caucasian, and the half were Asian; some were whole faces; some had the external features masked (hair) and others had the internal features masked (eyes, nose, mouth).\u00a0<\/p>\n\n\n\n When external features like hair were removed it often made it harder for participants to work out if a pair of faces were the same or two different people, especially when trying to match the faces of people from a different race to themselves. It suggests that external features are more useful in telling faces apart in these circumstances, Dr Havard said.<\/p>\n\n\n\n This was more prevalent with the Chinese participants when confronted with Caucasian faces with external features were removed, suggesting that Caucasian hair<\/em> might be an important cue for unfamiliar face matching tasks. This may be because Caucasian hair by its nature varies in colour and texture and therefore gives a useful cue, said Dr Havard, although she said hair was still used as a cue by the Caucasian participants in the task.<\/p>\n\n\n\n The study concluded that removing external features exacerbated what is known as the own-race bias<\/em> (ORB) for both groups, especially for Chinese participants matching Caucasian faces.<\/p>\n\n\n\n Dr Havard explained that existing research reveals that matching a face to a photo or deciding if two photos are the same person becomes even more error prone when the person making the decision is a different race to the faces they are trying to identify.<\/p>\n\n\n\n \u201cThis is called the own race bias (ORB), or cross race effect, a phenomenon whereby people are much better at identifying faces that are the same race as them, and much less accurate with faces that are of a different race to them,\u201d <\/em>she said.<\/p>\n\n\n\n The results demonstrated the importance of viewing whole faces when matching other-race faces.<\/p>\n\n\n\n Dr Havard said:<\/p>\n \u201cThis research has real world implications when it comes to tasks like matching someone to their passport or trying to identify those captured on CCTV to police custody suite images.<\/em><\/p>\n\n\n\n \u201cThis research demonstrates that when matching faces, more attention should be paid to the internal features of a person\u2019s face that cannot easily be altered, than the external features such as hair that can be altered or covered up.<\/em><\/p>\n\n\n\n \u201cIt also demonstrates that we should have people from diverse ethnic backgrounds when face identification is crucial for a decision e.g. at passport control or police trying to match criminals to CCTV footage.\u201d<\/em><\/p>\n<\/blockquote>\n\n\n\n Dr Havard\u2019s paper adds to her existing body of research work concerning own race bias in eyewitness identification and further work concerning face recognition.<\/p>\n\n\n\n It forms part of the OU\u2019s wider Forensic Cognition Research Group<\/a>, which works to share the research knowledge with police practitioners through the Centre of Police Research and Learning<\/a> as well as with other academics in psychology.<\/p>\n\n\n\n\n
Matching faces which are unfamiliar<\/strong><\/h2>\n\n\n\n
\u201cReal world implications\u201d<\/strong><\/h2>\n\n\n\n
\n
FIND OUT MORE<\/strong><\/h2>\n\n\n\n