From today, I will post more on my research topic, the human face and the meaning of human facial expression. In each post on this topic, I will contribute a philosophical point, and explore it in some detail with particular reference to Wittgenstein. Comments are always very welcome.
Can ‘facial recognition’ help Madeleine McCann police analyse picture? by Jonathan Richards (Times Online, September 26, 2007)
A picture taken by Spanish tourist Clara Torres, in north Morocco, which she believes shows Madeleine McCann.
Computer analysis of a picture allegedly taken of Madeleine McCann in Morocco is unlikely to provide more evidence that it is her than simply looking at it, experts said today.
So-called ‘facial recognition’ technology – which analyses an image of a face and seeks to match it with existing images of the subject – was not capable of making a more accurate assessment of a photograph than a human.
At best, all that a computer might conclude was that “this was a photo of a person who resembles Madeleine McCann a bit,” Dr Simon Prince, a senior lecturer in the Department of Computer Science at University College London, said.
The quality of the image was not even good enough to allow for a match – or what referred to as a “facial verification” – using the criteria that are typically applied in the analysis of, say, CCTV footage, Dr Prince said.
The number of ‘pixels’ between the centres of the girl’s eyes – a standard measurement in facial recognition – was 10, whereas usually 20 would be the “bare minimum” required in order for a computer to determine that two subjects matched.
“A typical digital camera might have 5 megapixels, meaning there are 5 million pixels in each image,” Dr Prince said. “In this photo, the area covered by Madeleine’s face makes up about 200 pixels, of which 10 are between her eyes, and that’s not enough. In addition to the resolution being too low, the face is partially occluded by the shadow coming from the woman’s arm, which further hampers identification.”
Facial recognition technology was principally used to search vast databases of pictures – for instance CCTV footage or the internet – that would take humans far longer to examine, he said.
The Child Exploitation and Online Protection (CEOP) Centre uses facial recognition to search through the thousands of photos that are seized in child sex offender investigations, and some cameras now also have facial recognition applications built-in which allow customers to organise their photo collections.
“The technology works by taking various reference points from a subject’s face – the distance between features, for instance – and then comparing those with the same measurements on different pictures,” a CEOP spokeswom said.
One common application is in building security systems, which can – by taking a picture – confirm that a person attempting to enter a certain area is who they say they are.
“Facial recognition works well when it is comparing passport-style photographs where the subject is front-on, the lighting is consistent, and the facial expressions are similar,” Dr Prince said.
“There can be problems, however, when the pose of the subject, as well as the lighting and expression, are different – that’s where all current research in the area is focused.”
Is recognizing a human face like computer analysis? Does it involve a comparison of two impressions, the current perceptual impression and an impression from prior experience? When the two match – “facial verification” – then one recognizes the face one is presently observing?
There are two conceptions at play here: a conception of memory, and a conception of recognition. First, it seems that memory preserves a picture of what has been seen before (PI, 604). This enables comparison of the perceived face with a picture of the face in one’s memory. Second, it appears that recognition consists in a fitting of the perceived face with the previously stored one. In this post, I offer two comments on the second conception. More to follow in future posts.
1. Need facial recognition involve a mental image of what is recognized?
Ordinarily not. It is atypical that people have memory-images when they recongize someone. Sometimes, a memory-image occurs after one has recognized someone. Or, the image may be of someone else associated with the person one is currently perceiving. Rather, the conception of recognition with fitting a picture to what it represents is applicable when “I look at the act of recognition (individual recognition) after the event; and not so much when I look at it to see what actually happened, as when I look at it through a preconceived schema” (PG, 182).
2. Is recognizing comparing what one perceives with a picture of it?
It seems that this metaphor leads to a regress. Even supposing that one’s memory-image and the perceived face fit, one needs to recognize the relation of image and face as a relation of fitting or agreement. One needs therefore to have a memory-image of the fit between the image and the face, and to judge that the image of the fitting fitted with the fitting. Does this even make sense? Is it an essential component of every case of facial recognition? It seems not (BB, 88).
More analysis on this topic to follow later this week and next.