The Quest In Biometrics
In the world of Biometrics, one of the crowing achievements of any R & D scientist, laboratory, or even any Biometrics Vendor is to have that particular piece of robust technology, or especially that really great algorithm. But what would make this technology or algorithm so great? Well, putting the FAR and FRR and ATV metrics aside, it would be to have extract the most unique features, either from a physiological standpoint, or behavioral standpoint. Really, all Biometrics currently extract the unique features from us, whether it is the shape of our hand, the minutae of our fingerprint, or even the way we type on a keyboard.
But, keep in mind, having that extremely robust algorithm is also a mathematical function of the uniqueness of the specific Biometric in question. For example, imagine a spectrum, and on it, imagine the gamut of Biometric Technologies which are available today. At one end of the least unique would be Hand Geometry Recognition, and at the other end would be Iris and Retinal Recognition. By far, the latter is the most unique and rich in terms of Biometric information and data. And if you really dig into much further, it is the retina which possesses the most unique and rich information.
And although I am far from being an expert on the eye, I have written a pretty exhaustive article on retinal recognition, which took many months of research and writing, and a couple of years ago, was published out of Europe. Many a Biometric Vendor has tried to capitalize on technology focused around the iris and the retina, some of have been really successful, and some have not been so successful, with creating that robust algorithm. The first true mathematical algorithms developed were that of Iris Recognition, by Dr. John Daugmann, of the University of the Cambridge, from the United Kingdom. His pioneering work led to the development of the technology which confirmed the before and after identities of that ever famous National Geographic picture of the Afghan girl known as “Sharbat Gula”. Probably of the of the other biggest advantages of iris and retinal recognition is that it is not prone, for the most part, from changes in the outside environment, and physiologically, it is very stable (versus, the finger, for example).
But probably the biggest impediment to iris and retinal recognition is that of privacy rights. People are just downright squeamish about having their eyes scanned, and heck, even I was when I first used an Iris Recognition device. But also, keep in mind, that in iris and retinal recognition, it is only certain components which are measured in that particular itself. For example, with the retina, it is the mapping of blood vessels in the back of the eye which leads into the optic nerve of the brain, which is examined.
With regards to the iris, it is the vector orientation of the unique features in the iris which is closely looked at for extraction. But, that is until now. Just today, I came across a brand new whitepaper released by the National Institutes of Standards and Technology (also known as “NIST”) which examines the other, surrounding parts of the eye as part of the overall authentication process.
This is known as “Ocular Recognition”, and here are some details of it: “Ocular recognition is a multi-components method which makes use of a variety of features in and around the human eye: eyebrows, eyelashes, eyelids, eye shape, sclera, iris, pupil, etc. Iris recognition makes use of some of the same multiple components, but only as a means to an end—to navigate and to extract the iris itself, from which all subsequent analytic comparisons are done. On the other hand, ocular recognition treats these components as features onto themselves which become part of the comparative process.” (SOURCE: http://www.nist.gov/customcf/get_pdf.cfm?pub_id=909790). And now, the researchers and scientists whom were involved in the compilation and publication of this white paper want to take this new part of iris and retinal recognition, and see how this new technique works under a broad range of applications, something which has never been done before.
Here are some more details of this: “With more factors at its disposal, ocular recognition hasthe potential for higher comparative power—bringing that potential to a practical reality is the challenge that this paper addresses . . . The broad goal of this paper is to encourage the development of robust ocular/iris recognition algorithms that yield accurate biometric results for a broad range of image and environmental conditions; e.g., a near-infrared (NIR) video- frame captured at a distance with an Iris On the Move (IOM) 1 system which we call “distantimage- based” contrasted with a classical image acquisition system . . .” (SOURCE: http://www.nist.gov/customcf/get_pdf.cfm?pub_id=909790).
The entire white paper can be downloaded here:
http://www.nist.gov/customcf/get_pdf.cfm?pub_id=909790
I plan to read this exciting new breakthrough, and of course, provide my viewpoints and commentary into a separate posting in the near future.
Until then, happy reading of this white paper!!!
Comments