These days any type of identification is beneficial. Forget fingerprints, retinal scans or the color of your eyes, and other markers of identity, security could soon be looking at the shape of your ears when deciding whether you are who you say you are. This week’s Hearing International looks at ear identity scans. 

Researchers have discovered that each person’s ears have a unique shape and are as distinguishing as your fingerprints; no two ears, even on the same person, are alike. Ears can be scanned and then be compared with a database of ear shapes to identify people. While one might think that these ear scanning systems have just been discovered, they have actually been around for quite some time. In fact, the first patent for the use of the outer ear for identification purposes was granted to Manual Zimberhoff in 1963.  Prior to and subsequently lots of domestic and international research has been conducted in this area. 

One system that takes advantage of previous research has been offered by Abaza, Hebert & Harrison (2010). 

Their system, using shape-finding algorithm called the “image ray transform”, is thought to be 99.6 % accurate and could make the outer ear into one of the most accurate and least intrusive ways to identify people.  Abaza et al (2010) describes their system in three components, Pre Processing, Image Normalization and Edge Segmentation and Localization. First, the Pre Processing segment involves detection and segmentation of the ear region from the overall image,  while the Image Normalization component normalizes the intensity of the ear region and aligns the ear. The Edge and Localization component finds the external contour within a segmented region of the outer ear. Earlier geometrical measurement work by Iannarelli (1989) is used feature extraction representing the various ear structures by features. Classifications matching the probe and the gallery feature vectors are then used to identify the specific ear. In the verification mode, the subject claims an identity and the system verifies that claim. Later, in the identification mode, the system searches a database for the identity of the scanned individual.  Professor Mark Nixon, who led the team from the school of electronics and computer science at the University of Southampton, said: “There are a whole load of structures in the ear that you can use to get a set of measurements that are unique to an individual.

While this process could be used in the future for passport control to identify individuals as they simply walk through immigration, the system is now available to various security groups and law enforcement. One startup company, Helix Biometric Ear Recognition Software has developed a system that works with cameras, cell phone cameras, and police body cameras to recognize individuals. While  not an endorsement of the Helix system, their success does present the development of this process and how far it has come in a rather short time. Click on the Helix picture or here for an introductory video of their system and how its used.  Descartes Biometrics has also come up with an app that can identify smartphone users by the way they press their phone to their ear and cheek—though its less-than-consistent recognition means that perhaps this particular app isn’t quite ready for prime time.

Looks like the future is here for ear recognition; George Orwell’s omnipresent government surveillance presented in 1984 is one step closer to reality.  

 

 

References:

Abaza, A, Hebert, C. & Harrison, M. (2010).  Fast learning ear detection for real-time surveillance.  Presentation to the fourth IEEE Biometrics, September 2010.  Retrieved October 9, 2017.

Gray, R. (2010). Ears provide new way of identifying people in airports.  The Telegraph  Retrieved October 8, 2017.

Iannarelli, A. (1989). Ear Identification, Forensic Identification Series, Paramont Publishing
Company, Fremont, California, 1989.  Retrieved October 10, 2017. 

Zimberoff, M. (1963).  US patent US3102459 A.  Photographic ear identification system.  Retrieved October 8, 2017.

Leave a Reply

Your email address will not be published.