Mark S. Nixon, professor in computer vision at the School of Electronics and Computer Science, University of Southampton, UK, loves recounting a favorite Sherlock Holmes story.
In "The Disappearance of Lady Frances Carfax," a distinctively shaped ear gives a suspect away.
Life imitates art as Nixon pioneers biometrics research using the ear, rather than the face or other features, for identification purposes.
Using photos of individual ears matched against a comparative database, Nixon and other researchers believe they have a means of identification as distinctive as fingerprints.
Though ear identification has yet to find use in any commercial application, security will be the best use of the technology, Nixon said in a phone interview last week
In a March 2010 paper, “A Survey on Ear Biometrics,” Nixon and researchers Ayman Abaza, Christina Herbert, and Mary Ann F. Harrison of the West Virginia High Tech Consortium Foundation and Arun Ross of West Virginia University write:
Humans have used body characteristics such as face and voice for thousands of years to recognize each other. In contemporary society, there is a pronounced interest in developing machine recognition systems that can be used for automated human recognition. With applications ranging from forensics to national security, biometrics is slowly becoming an integral part of modern society.
Using ears for identification has clear advantages over other kinds of biometric identifications, Nixon said.
Once developed, the ear changes little throughout a person’s life, providing a cradle-to-grave method of identification, he explained.
Ear identification also could offer an important, non-contact alternative to fingerprints or retina scan. During walk-throughs at security checkpoints, cameras could digitally photograph passersby, comparing their ears against others in a database.
Even among other non-contact biometrics, ear identification offers key advantages. Used in combination with face recognition, ear recognition offers a second point of comparison in cases where all or part of a face might be obstructed, for example, by makeup or some other alteration.
Ear identification trumps other, more-intrusive approaches, Nixon told me. Keeping images of ears, as opposed to faces for comparisons, may raise fewer privacy concerns, he believes.
Watch a presentation from Nixon on the History of Biometrics at the Institute of Electrical and Electronics Engineers International Joint Conference on Biometrics in the US in 2011.
What do you think of the use of ear identification or recognition for security applications? Leave your comments below.
The single most important thing that could be done to improve the biometrics discourse would be to re-frame the language used. There can be no such thing as a "perfect" biometric, and I don't mean that in the philosophical sense.
Biometrics always commit two types of error: False Accepts and False Rejects. Because real peoples' biological traits change over time, and because they present differently each time, there's an inevitable tradeoff. The lower the False Accept rate, the higher the False Reject rate, and vice versa.
So what could "perfect" mean? A biometric that never ever confused you for someone else would refuse to recognise you. And a biometric that never ever rejected you, would happily accept others instead of you.
Shawn, am first time hearing about identifying ear as a part of biometric application. So far we had seen about face reorganization, thump impression and retinal analysis for biometric identification. But I don't know whether the ear patters are either unique or different across peoples. But as you mentioned it has an advantage of no need for physical contact with the instrument, so hopefully it's more hygienic.
Thank you Shawn, from that it sounds like unless you make drastic changes to your ear, or have an injury that the system would work for quite a long time. I suppose after someone hits 70 years of age they can just re-scan every few years the way we renew driver's licenses to keep up with those age related changes.
Just for your and everyone else's edification, I thought I'd add some excerpts from the study to which Mark Nixon contributed, "A Survey on Ear Biometrics," for more on how ear recognition is done:
[The] ear biometric system may be viewed as a typical pattern recognition system where the input image is reduced to a set of features that is subsequently used to compare against the feature sets of other images in order to determine its identity. Ear recognition can be accomplished using 2D images of the ear or 3D point clouds that capture the three-dimensional details of the ear surface.
Here's more on the change of ears overtime:
The forensic science literature reports that ear growth after the first four months of birth is highly linear. The rate of stretching is approximately five times greater than normal during the period from four months to the age of eight; after which it is constant until around the age of seventy when it again increases.
Beth, that is what I was thinking would be the most common application for that kind of biometric platform.
Think of how it would apply in London, with all the survalence they have. just randomly scanning ears as they pop up on video and run it against a criminal database. I dont think something like that would really fly in the US, but in England they already have the network and lack of privacy laws to make it happen.
Shawn. As someone who has just gone through the fiasco having to get legible fingerprints for a security clearance, I can tell you this is welcome news. Anything has to be better from the user experience. I had both ink and digital and they were never quite good enough. The whole experience of having to go to a police station is very uncomfortable. Would't it be nice to just say..let me show you my ear!