Tag Archives: Face Recognition Technology

Was a facial recognition technology failure behind yesterday’s Australian Border Force debacle at Australian airports?

https://www.sbs.com.au/news/was-australia-at-risk-while-automated-immigration-systems-were-down-at-airports-across-the-country

Not so SmartGate!

Doppelganger doco on Nine tonight

https://www.yourtv.com.au/program/finding-my-twin-stranger/358709/

My thoughts after viewing:

Not the most entertaining or interesting documentary that I’ve seen.

The pairs of “doppelgangers” featured on the show highlight the difference between popular or conventional notions of a doppelganger and the concept as I use it, being a super-recognizer. I didn’t find it clear how the pairs of supposedly same-looking people had been discovered, but I think some pairs discovered each other by accident by living in the same town or doing stand-up comedy  in the same festival and being mistaken for the other, while other pairs used some online face recognition tool to identify their “double” in another part of the world. In the matching methods, therefore, some pairs were matched by people looking at live, dynamic people, while other matches were made by technology looking at still images of faces. Not surprisingly, one pair who were matched with face shots and technology were of very different heights in person and not much of a match for skin colour or general facial resemblance.

All of the pairs on the show were matched in genders, age, skin colour, hair colour, hair texture and even non-biological aspects of appearance such the use of glasses to see with and even similar style of glasses worn. Throughout the show the scientist studying the pairs, using computer technology to compare static images or 3D computer models of the pairs faces, compared the similarity scores given for the matches to those typically found in identical twins, so the unspoken concept of the doppelganger used in the show was for the pairs to look so similar as to be identical, but none of the pairs had similarity scores in the same range as identical twins, overall. As soon as our daughter and I started watching the show we could pick that none were identical twins; there were always differences in faces that could be spotted in an instant that wouldn’t be there in identical twins. This concept of a twin-like doppelganger is entirely different to the kind of uncanny similarity that I occassionally spot between people, which is much more like a family or genetic similarity, as in close relatives or people who have the same genetic disorder that alterns appearance, but often neither of these explaantions are obvious. The doppelganger phenomenon that I spot can go across ages, genders, races and skin tones, but often the personalities are uncannily similar, in the same way that their faces are similar. My concept of “doppelganger” violates social norms, in that it suggests that there are more fundamental similarities between people than sharing the same gender or race or age, which many people might find odd or insulting.

The resemblance that I sometimes see between apparent strangers is in multiple aspects of the face that are remarkably similar in shape and appearance, which can include the hairline, the texture of the hair, beard-borderline pattern, the pattern of the teeth and jaw width, along with things that can’t be recorded in a static photograph, such as the way a person speaks, pronounces particular phonemes in a way that is independent of an accent, the rhythm and speed of the way they speak, unique or distinctive facial expressions and the context in which they are made, which might seen incongruous, and similar gestures or postures, such as the angle at which the head is usually held, and the overall personality. My concept of “doppelgangers” goes way beyond simple visual matching of two similar but non-identical images of faces, which is a task that even a machine could be designed to do. My concept of the doppelganger takes in the whole package of sound, speech, movement and facial appearance, and none of those elements alone are interesting or remarkable, because it is the matching of the same convergence of these types of characteristics in two different people that I notice. I believe this is a reflection of a biological similarity between people, and I think there was a hint of a similar sense of biological similarity at work in the documentary. Part-way through the doco pairs were shown givng saliva samples so that they could be compared for genetic similarities, in a similar process to the popular geneological services that aim to identify distant relatives by DNA. While one pair had remarkably similar racial profiles, it appears that only one (other) pair turned out to be actually related, and they were the pair that I felt I’d have the most trouble mixing up if I met them both, because of their similar overall movement style, voices and personalities and appearance. They were both stand-up comedians, which must count as a peronality similarity. This pair nevertheless did not receive a high score from the computer for facial similarity, but tellingly, they did receive the highest similarity score when rated by a crowd of people. Sadly, the fact that human rating managed to identify an objectively real genetic similarity in one of the pairs appeared to be ignored, in the documentary while the similarity ratings of static images by a computer algorithm was spoken of as an objective fact. Once again, it appears that the common infatuation with and awe of technology is a barrier to expanding scientific knowedge of identfying other people.

Nine News on facial recognition technology and our surveillance state

https://www.9news.com.au/2019/02/10/20/31/news-australia-facial-recognition-passport-federal-government

 

Super-recognizers on Australian public radio today

 

Genelle Weule So, you think you’re good at recognising faces. ABC Science. March 11 2018.

http://www.abc.net.au/news/science/2018-03-11/super-face-recognisers-are-you-one/95177

Super-recognisers. All in the Mind. ABC Radio National. 11 March 2018.

http://www.abc.net.au/radionational/programs/allinthemind/super-recognisers/9523296

Another facial recognition tech fail

https://www.theguardian.com/technology/2017/sep/14/apple-face-id-didnt-fail-at-iphone-x-launch-staff

 

Facial recognition tech fail

Police facial recognition trial led to erroneous arrest
Alexander J Martin
Sky News August 31st 2017
http://news.sky.com/story/police-facial-recognition-trial-led-to-erroneous-arrest-11013418

 

Warning from a counter-terrorism expert

Australian airport security lack key skill to protect travellers: expert.

James Ried
The New Daily.
http://thenewdaily.com.au/news/national/2017/07/31/australian-airport-security-lack-skill-expert/

 

Mind the hype about AI and machine learning

I’m an advocate of the use and hiring of super-recognizers in various work roles and tasks that involve face memory or face recognition, but there are lots of people, some with personal interests and some without, who believe that technology can or will do a better job than humans can. The interesting articles below are probably worth a read if you have an interest in new technology relevant to face processing.

Hooker, Giles Machine learning: What journalists need to know. Sense About Science USA. March 22 2017.

http://www.senseaboutscienceusa.org/machine-learning-journalists-need-know/

Reese, Hope Top 10 AI failures of 2016. TechRepublic. December 2, 2016.

http://www.techrepublic.com/article/top-10-ai-failures-of-2016

 

Australia’s Face Verification Service

News that I didn’t have time to write about in November last year. The federal government of Australia plans to create a database of photo-id images. I’m not sure if this is just the National Facial Biometric Matching Capability renamed.

“A Face Identification Service (FIS) is expected to commence in 2017 to help determine the identity of unknown persons. It will be used for investigations of serious offences by specialist officers.”

Couldn’t be less detail or description or explanation in this statement.

https://www.ag.gov.au/RightsAndProtections/IdentitySecurity/Pages/Face-verification-service.aspx

https://www.itnews.com.au/news/australias-new-facial-verification-system-goes-live-441484

http://www.zdnet.com/article/australian-face-verification-service-starts-with-citizenship-imagery/

 

Interesting questions and serious concerns

Revell, Timothy Concerns as face recognition tech used to ‘identify’ criminals. New Scientist. December 1st 2016.

https://www.newscientist.com/article/2114900-concerns-as-face-recognition-tech-used-to-identify-criminals/

Garvie, Clare, Bedoya, Alvaro and Frankle, Jonathan The perpetual line-up: unregulated police face recognition in America. Center on Privacy & Technology at Georgetown Law. OCTOBER 18th 2016.

https://www.perpetuallineup.org/

Is there really a criminal face? I don’t think the research discussed in the New Scientist article settles the debate by any means, but at least the controversial idea is opened up for investigation. If there is one my guess is that it is a look that coincides with the Australian face (every race and nation has a distinctive averaged facial type, apparently). European colonisation of Australia began as a penal colony and thus a good part of the white genetics of Australians arrived in this country through people identified as criminals. My best guess is that the crim face has a large straight nose, thin lips and puffy, small eyes. I’d guess this unattractive face could in itself be a social and economic disadvantage, or could be symptomatic of a phenotype that includes some degree of intellectual impairment. I think if there is a crim face it might have little to do with personality but a lot to do with disadvantage, but this is all speculation.

I think it is worth noting that claims made in the print version of this article about supposed advantages of AI over humans in face recognition skills such as identifying age, gender, ethnicity and tiredness by looking at faces presumably only apply to humans of average face recognition ability who maybe are not as exhaustively trained in these skills as the AI systems have been. One cannot compare human ability with AI in face recognition until appropriately trained super-recognizers (representing the top end of human ability) have been pitted against machines. I’m guessing this hasn’t been done.

Perhaps the most important part of this article is right at the end; “…the majority of US police departments using face recognition do little to ensure that the software is accurate.” That certainly is not good enough. Human super-recognizers have abilities that have been proven in scientific testing and also in practice in policing in the UK. Why do so many people persist in the assumption that machines must be better than humans in visual processing, in the face of an abundance of evidence? The link in the New Scientist article to the website of the researchers who have criticized the use of face recognition technology in law enforcement in the United States of America is worth a look for sure.