What the Nose Knows
IN RUSSIAN WRITER Nikolai Gogol’s short story “The Nose,” a civil servant’s nose leaves his face and tries to establish its own identity. In the real world, a nose has not been examined as a standalone biometric feature that could be used to identify individuals. It has, instead, mainly been considered in conjunction with the entire face, says biometric expert Arun Ross of the computer science department at West Virginia University.
That could change. Adrian Evans, lecturer in the department of electronic and electrical engineering at the University of Bath in the United Kingdom, and coresearcher Adrian Moorhouse (also Evans’ student) are taking a look at the ability of computers to recognize a person based on the nose alone. They chose the nose, in part, because Evans says it’s difficult to discreetly hide one’s nose, and the nose isn’t affected by changes in facial expression.
The researchers enlisted 36 volunteers and scanned their faces multiple times using a software program called PhotoFace, developed at the University of the West of England. PhotoFace uses two cameras to generate a three-dimensional image, according to Evans.
Ross says that the research is particularly interesting for two reasons: “One, they are using photometric stereo to obtain a three-dimensional image of a person’s face, and subsequently the person’s nose. And two, they’re trying a novel biometric in terms of just using the features of the nose.”
Evans and Moorhouse spotlighted six different types of noses: Greek, Roman, Nubian, snub, turned-up, and hawk. The areas on the nose that were isolated are the ridge, the tip, and the nasion, which is the space between the eyebrows. Evans says that the researchers found that the ridge line between these points is a good one-dimensional measure that can distinguish between various noses. He says it is easy and quick to capture, including on unwilling participants. Additionally, ratios between the various points on the nose were used to help identify subjects.
Evans thinks nose biometrics could be useful at building entrances, where one could have a quick photo taken for the nasal biometric instead of having to swipe a card or place a finger on a fingerprint pad. Additionally, he says nose biometrics could possibly be used in airports from surveillance footage.
Evans and Moorhouse are not the only ones researching the potential of a nose biometric. Researchers Hassen Drira, Mohamed Daoubi, and Boulbaba Ben Amor from the Université de Lille in France and Anuj Srivastava from Florida State University have recently published papers examining the success of various aspects of the nose in identification.
Not everyone is on board with the idea of using the nose as a solo biometric, however. Detractors point out that rhinoplasty, an accident, or even age can change the shape of the nose, and that face scarves would obviously pose a challenge.
Research indicated that some issues to be considered are that full nose biometrics are computationally expensive and absolute matches are harder to achieve than with other biometrics, according to a presentation the researchers gave at the recent International Conference on Imaging for Crime Detection and Prevention. “It’s definitely fair to say that the nose wouldn’t compete against a full iris system at the moment,” says Evans.
However, he adds, “because it’s easy to capture, the way we see it going might be in conjunction [with other identity biometrics]. So you could use noses, perhaps, to cut down a database to a [smaller] number of likely matches.” Ross agrees that that method, known as indexing, might be a helpful use for the nose in biometrics. “Let’s assume you have a database of 10,000 subjects. By using the nose alone, you can maybe prune down the number of possible matches to say about a hundred or 200… that would be a very positive outcome if the authors are able to demonstrate that you can use the nose to reduce the number of possible matches in very quick time so that you don’t have to go through the entire database exhaustively.” However, Ross adds that work must be done to test the capture of noses under different conditions, such as in poorer lighting and using different poses going by at different speeds.
Another issue is that the “work to see how unique a nose is” has not been done yet, says Evans. For example, with irises, it is known that even identical twins will have different irises, says Evans. Nasal dimension uniqueness is not as clear yet.