Tag Archives: Person-recognition system

Person recognition a crucial point in plotline in another classic movie?

We stayed up late to watch the 1940s classic movie Gaslight the other night. What an amazing story and brilliant acting from Boyer and Bergman! Charles Boyer was not an unusually handsome man, but her certainly knew how to use his eyes, and he also knew how to use his voice.

Was it the detective recognizing a family resemblance between Paula and her late aunt that caught the attention of the detective, or was it that combined with a visual recognition of Serges Bauer?



Interesting letter from last year re acquired prosopagnosia and the uncanny valley

Editor’s pick: Excluded from the uncanny valley
From Bob Cockshott

New Scientist. Issue 3100. November 19th 2016. p.60.


The experience described in the letter seems to suggest that there is more to face recognition than simple memory or faces, or could it be that there are aspects of the perception of faces in particular that make them especially memorable? Faces are easy to personify because they are usually found on people. Perhaps it is the ability to detect the person behind the face that has become amiss in the letter’s author following his stroke, and maybe this ability feeds into face memory? Such a relationship would explain the author’s inability to notice the uncanny valley, and it would also explain why a personifying synaesthete like myself is also a super-recognizer.



I’ve had a read of the interesting article by Laura Spinney that this letter was a comment about, and I think the Perceptual Mismatch Theory of the Uncanny Valley Effect probably offers a more plausible explanation of my a prosopagnosic might be unable to detect the uncanny valley than the competing Category Uncertainty Theory. The article explained the two theories and evidence supporting them. To summarise, the CUT explains the UVE as the result of confusion about what type of thing one is looking at (for example robot or human?), while the PMT explains the UVE as resulting from unease or perceptual confusion when different features or parts of the thing or being viewed have dissimilar levels of human-like appearance (for example the face and skin look realistic but the eyes do not move like human eyes). I think the case of a prosopagnosic not detecting the UVE when people with normal face perception do is support for the PMT theory rather than the other because as far as I know, prosopagnosia does not involve inability to classify faces or bodies as human or non-human, while I believe there is evidence supporting the idea that prosopagnosia can be the result of not being able to perceptually integrate the features of the face as a whole that is recognizable as a unique or distinctive mix of many attributes and features. A non-prosopagnosic person should be able to perceive a face or body as a whole made up of parts, and notice if one or more elements has a level of humanity that does not match other parts, as in the PMT, while a prosopagnosic might not. Of course, research is needed to investigate my armchair speculations.

It makes sense to keep track of troublesome humans

Taub, Ben Antarctic Birds Can Recognize And Hold Grudges Against Specific Humans. IFLS. March 28th 2016.



Won Young Lee, Yeong-Deok Han, Sang-im Lee, Piotr G. Jablonski, Jin-Woo Jung, Jeong-Hoon Kim Antarctic skuas recognize individual humans. Animal Cognition. First online: 3rd March 2016.



They are developing tests for recruiting super-recognizers into the police in the UK, but don’t ask me what’s happening here

Phillips, Mark London police using crime-fighting “super recognizers” official. CBS News, Dailymotion. Publications date November 12th 2013. http://www.dailymotion.com/video/x173o5e_london-police-using-crime-fighting-super-recognizers_news

This is an American report from CBS News published in November 2013 on the use of super-recognizers in London policing. Detective Chief Inspector Mick Neville explains how inadequate computer facial recognition was found to be compared with results from police supers. PC Gary Collins and researcher Dr Josh Davis are also interviewed, and super-recognizer police doing identification work are shown. It is revealed that tests are being developed for recruiting super-recognizers into a police force in London. My guess is that this would mean recruiting supers into the police force because they are supers, in addition to their existing policy of finding and utilizing the many supers that they already have serving in this large police force.

Barone, Tayissa Council’s eyes guide long arm of the law. West Australian. September 7th-8th 3013, p.20-21 news.

“The Met” continue to be leaders in the use of human facial super-recognition in policing, but what is happening here in Western Australia with regard to human face recognition and CCTV and policing? As far as I can tell, not a lot. As far as I know there is no testing of any kind of face recognition or face memory ability in police recruitment, and I’ve not read anything about use of supers in any Australian police force. In September 2013 the Weekend West had an article in it about operators at the City of Perth’s surveillance centre working with and beside members of the WA Police to keep things under control in the city streets. The journalist wrote about the tens of millions of dollars that the City of Perth has spent on their CCTV camera network, the “unique” skill set of the surveillance centre operators, their intuitive understanding of body language, their eye for detail, multitasking ability, the keen competition for their jobs and some rigorous battery of testing in which only one out of 160 applicants met the required standard, but not a single mention of face recognition or visual memory.

Western Australia Police Service reduces crime through intelligence-led policing with ABM. ABM United Kingdom Limited. 2012. http://www.abmsoftware.com/en-GB/products/82-uk/news/case-studies/139-western-australia-police-service-reduces-crime-through-intelligence-led-policing-with-abm.html


A webpage of a software company ABM boasts that it provides the WA Police with facial recognition technology for use on photos and other static images of offenders, which will probably impress the “boys who love toys” technophile set, but it fails to impress me for a couple of reasons. Firstly, there appears to be good evidence that the most able humans can vastly out-perform contemporary facial recognition technology, and secondly, the identification of people from still images of faces or entire bodies is inherently limited. It does not use the wealth of information that one can glean from looking at a moving image. Faces are unique and so are the ways that people move. A moving image is essentially richer and more complete and more natural than looking at a single still image. Study of interviews of super-recognizers yields many clues that supers recognize people, not images and not just faces. Supers can identify people from photos, but it seems likely that the memorization process works best if it is based on watching people, not looking at photos. I am not aware of any face recognition technology that works off moving images, but that might just be a mark of my ignorance. I remain skeptical.

The idea that technology must beat humans in face recognition is a popular one, I think based on some major misconceptions about human psychology and artificial intelligence. I think a lot of people assume that if tasks like visual identification or walking or recognizing voices are effortless for humans then they must be even more easy for a computer system to perform. This shows an ignorance of the millions of years of biological evolution that gave humans and even the most humble animals sensory perception, and the sensory and movement systems of muscles and nerves that give rise to the power of voluntary movement. These processes involve brains as much as they involve sensory organs and muscles. The fact that we are able to do these things without thinking much about them is no indication at all that they are simple. It is just an indication that some of the really clever tasks in cognition are too complex and important to be exposed to the interference of conscious thinking. Attempting to recapitulate the kind of design complexity that is found in biological sensory perception and biological movement with technology and computers would surely keep a designer occupied for a very long time. Good luck with that.

Doubts that legendary memory performer had any advice for prosopagnosics

On the weekend we visited “the hills”. Communities in picturesque forest locations a bit away from the city tend to have a lot of arts and crafts and historical attractions and also lots of second-hand book shops. They are a lot like Fremantle, but not a port and there’s less foreign tourists and probably not nearly as many junkies. Can any community have too much “arts and crafts”? I think it is possible. Why paint life badly when you could be experiencing it instead? Browsing second-hand bookshops is also a questionable use of time, but I can’t keep out of them. I know the vast majority of their stock has a value that is close to landfill, but now and then I come across a forgotten book that adds something unique to a current interest. In one second-hand bookshop I spotted an old, stained and overpriced copy of a book about recognizing people written by the legendary mnemonist Harry Lorayne, whose specialty was memorizing the names of people in the audience in incredible quantities, presumably mentally linking names with faces.

For a person with a scientific interest in face memory, a performer like Lorrayne is of interest. How did he do it? Did he have superior face memory? Did he have a technique for improving face memory? I didn’t buy the book by Lorrayne, but I scanned through it to get an idea what all of the chapters were about, and as far as I can tell, there’s no technique in the book except  mnemonic techniques for creating and memorizing visual images that are visual-conceptual mnemonics for linking people’s names to their faces. As far as I can tell Lorayne’s techniques are all about linking names with faces, but offer no tips or help in visually memorizing the faces. As far as I can tell, Lorrayne takes normal face memory in the reader for granted, so I doubt that a prosopagnosic would find anything to help in his book. I can imagine that a face-blind person might have bought this book in the hope that it would help, and be left disappointed and confused. We should be very grateful to the researchers in psychology and neuroscience who are giving us more and more real information and advice about face memory and prosopagnosia and other perceptual abilities and disabilities. Reliable information, useful tests and the latest research findings can be found through the internet. Ignorance should be left behind in dusty old second-hand book shops.

Fashion recognition app that could be useful for prosopagnosics

Google Glass app identifies you by your fashion sense. by Paul Marks New Scientist. Magazine issue 2907, 7 March 2013.

I’m not completely clear how this recognition technology works, but it says it creates a spatiogram that is a record of what a person is wearing, including colours, and it can be used to identify an individual in a crowd. I found it interesting that face recognition was dismissed as unfeasable, considering the many articles I’ve read over the years making big claims for face recognition technology:

“Face recognition systems cannot be used for this, says InSight developer Srihari Nelakuditi at the University of South Carolina in Columbia, because it is unlikely someone in a crowd will be looking straight at a headset’s camera.”

Interesting stuff from the New York Times from late last year about prosopagnosia and another condition that I’d not heard of – phonagnosia

Have We Met? Tracing Face Blindness to Its Roots by Karen Barrow New York Times December 26, 2011   http://www.nytimes.com/2011/12/27/health/views/face-and-voice-recognition-may-be-linked-in-the-brain-research-suggests.html?_r=2

this is the study mentioned in the above article:

Direct Structural Connections between Voice- and Face-Recognition Areas. Helen Blank, Alfred Anwander, and Katharina von Kriegstein Journal of Neuroscience. 7 September 2011,31(36): 12906-12915; doi: 10.1523/​JNEUROSCI.2091-11.2011  http://www.jneurosci.org/content/31/36/12906.abstract?sid=2704774c-0cbd-4342-9b75-639b388b99f9

and this is the video featuring Dori Frame that goes with the article:

Faceless produced by Almudena Toral New York Times December 2011 http://video.nytimes.com/video/2011/12/26/health/100000000958226/faceless.html