Talk:Iris recognition/Archive 1

Scanning driver's iris's at 50mph

I remember reading an article some years ago that said that British Telecomm had developed a system that could scan driver's irises in cars passing at 50mph. This may explain why irises are going to be used in the (horrible 1984-Nazi) UK national identify cards instead of digitised fingerprints, which one would otherwise have thought of as a better choice, as the government would like to track the movements of terrorists and other criminals.—The preceding unsigned comment was added by 81.104.12.74 (talk • contribs) 16:08, 17 May 2006 (UTC).

I try to follow the scientific literature related to iris-scanning closely and have not managed to find any original reference to such a system that details its actual performance. Can you provide a reference to anything published by the developers or independent evaluators (rather than writings of privacy activists or journalists who sometimes like to make their point with grossly exaggerated and unrealistic claims that later turn out to be merely based on distorted rumours or misunderstandings, e.g. [1])? It is difficult to see how one could implement a drive-by iris scanner with any useful false-reject rate. An actually available recognition system for moving people is for example Sarnoff's "Iris on the Move" system [2], were iris scans are performed while people walk through a gate. While such systems work in lab demonstrations, they have a much worse false reject rate than systems that rely on the cooperation of the scanned person. Markus Kuhn 10:53, 17 December 2006 (UTC)

I worked at British Telecom on Iris Scanning and cannot remember any of us making this claim. For iris scanning to be effective, you need a few hundred pixels across the iris, which would mean that a VGA video camera would need to be zoomed in on just the eyes. A high resolution camera would need to zoom in to frame a photo of just the face. Whilst I suppose this is just about possible with a paparazzi-style telephoto lens with image stabilization and ideal bright lighting, it would not be easy and so I think a moving target at 50mph in a dimly lit car through a windscreen would be just about impossible.

Patent Status

Is the patent mentioned in the first paragraph the one that recently expired? If so, that would be good information to make note of. In articles I had read, it was called the "Flom" patent. [Businessweek Article] 71.213.88.75 23:40, 24 October 2006 (UTC)

Iris cross-comparison figures

Someone has several times removed a reference to the best available estimates of the false-accept rate of Daugman's IrisCode, based on the cross-comparison of "200 billion iris pairs", with the argument that this can't possibly be true because these would be far more iris pairs than there are human individuals on Earth. I understand well that among some privacy activists it is not customary to actually read the relevant scientific literature (the article has the relevant link [3]) before rubishing its results, but in this case even reading the abstract would easily have resolved the misunderstanding. In order to get statistics from comparing 200 billion iris pairs, one needs images of only n = 632 500 irises (from a United Arab Emirates immigration database), which then enable C_n² = n · (n − 1) / 2 = 200 027 808 750 comparisons between pairs of irises enroled in this database. (See also birthday paradox.) Markus Kuhn 11:11, 17 December 2006 (UTC)

IrisCode patent

Is this the IrisCode patent application mentioned in the article: http://google.com/patents?id=KRkpAAAAEBAJ? Danny 21:32, 1 August 2007 (UTC)

"Deployed applications" section.

I'm proposing a bit of a cleanup of this section, which seems to have collected a lot of WP:SOAP and other cruft recently. Other views, before I make a start? --Old Moonraker (talk) 16:27, 9 January 2009 (UTC)

Some WP:COPYVIO from these press releases and some WP:SOAP removed. Please stop reinstating this copyright material. --Old Moonraker (talk) 11:15, 11 January 2009 (UTC)

Inline refs

The sections on advantages and disadvantages and the article in general are lacking inline citations.

I'm a bit confused - I added a link to the bottom of this page which discussed the wide variety of applicatioins that iris recognition has, and I was told that it went against guidelines and I have absolutely no idea why, as the link did not promote products at all. Can someone help me out on this? Here is the link: http://www.smartsensors.co.uk/es/wp-content/uploads/2010/10/Use_Cases_Jan_2011_website.pdf (Wanausha Khafaf (talk) 13:14, 28 April 2011 (UTC))

iris expansion

what is the iris expansion

WP:COI

Recent additions by User:Smhossei seem to refer to his own work: "He also introduced a new and alternative feature extraction method to encode VW iris images which are highly robust through reflectivity terms in iris. Such fusion results are seemed to be alternative approach for multi-modal biometric systems which intend to reach high accuracies of recognition in large databanks" and may contravene the WP:COI restrictions on editing and some WP:PEACOCK seems to have slipped in. The source, published under the auspices of the IEEE, seems to be sound but is this nonetheless just self-promotion?--Old Moonraker (talk) 15:04, 28 July 2010 (UTC)

what is the cost of iris scanner

What is the cost of iris scanner in india. —Preceding unsigned comment added by 117.204.25.93 (talk) 05:05, 17 December 2010 (UTC)

Robtics project attention needed

• Bullets to prose
• Check refs
• Prose
• reassess

Chaosdruid (talk) 10:14, 24 March 2012 (UTC)