CREDIT: Pixabay

Previously, it took biometrics such as facial recognition, fingerprints or retinal scans to identify the unique physical characteristics of each human, but now identification might be as simple as measuring your ear.

Using ears to identify humans isn’t new, but our favorite crime shows still rely on old faithful methods — dusting for fingerprints or collecting DNA — to land a suspect. In all fairness, criminals are much more likely to touch things with their hands than rub their ears all over a crime scene.

It was, however, used in a 1997 murder trial after an investigation lifted an ear-print from the window the killer entered through. The subsequent conviction was the first using ear prints as an identifier. However, the verdict in 2004 was overturned on appeal as DNA from the ear print indicated a different suspect. It was the opinion of the expert used in the trial that solidified the guilty verdict.

That case was flawed, but recent studies show ears are just as reliable an identifier as our fingerprints.

A team of researchers from the University of Georgia in 2022 developed software that scans your ear. It was intended to serve a post-COVID world in which people wear masks — muffling voice recognition — and are conscious of what they touch.

Masks aren’t the problem they used to be, but there are other security operations in which ear identification can be instrumental.

Ears are fully formed and developed at birth. Except for the consequences of age they really don’t change over time. Each ear is unique, and your ears are even unique from each other. This makes them a reliable source of identification – even from a distance.

You can’t access someone’s fingerprints or DNA from a photo, but even a photograph of your ear can tell us who you are. And with the number of crimes recorded on video, ear biometrics can help identify the culprits.

A more recent development in ear identification came when a team of forensic and dental scientists from all over the world built on a 2011 study by Roberto Cameriere that measured the four anatomic regions of the ear and combined the measurements to produce a code that is unique to each person.

They implemented a larger specimen group and divided it across multiple ethnic groups to stretch the method and determine further accuracy. They found that when they added the codes for each person’s ears together, there were zero code repeats. This means 814 unique ear identifiers. The team concluded that “the probability of two different individuals having the same code (false-positive identification) was found to be less than .07 percent.”

So, if you’re planning to launch a crime wave, make sure to wear ear muffs. They’ll protect you from the cold and from getting caught.

But if you happen to forget, perhaps you can simply slouch your way through your criminal activity. Or not.

A team of researchers from Khalifa University suggest the factors that inhibit accurate ear identification in 2D and 3D images — posture, light and scaling — can be overcome with combining both, “To the best of our knowledge, this is the first time two-dimensional and three-dimensional ear attributes have been merged to build a detector and descriptor for matching a pair of 3D ears. Combining features from the 2D domain and features from the 3D domain considerably increased recognition efficiency.”

The team suggests that a keypoint detector and a descriptor, built from angular features of 2D ears and textures of 3D ears, can lead to more accurate ear identification. The texture and shape combined enhance the veracity of the results.

“This holistic approach culminates in the achievement of state-of-the-art results while simultaneously ensuring robustness to illumination and pose variations,” says Iyyakutti Iyappan Ganapathi. He is lead author on the study and a post-doctoral fellow in the electrical engineering and computer science department at Khalifa University.

Ganapathi says while there is comparable accuracy between other commonly used biometrics and ear identification, a lack of data is a challenge.

However, he is hopeful going forward.

“Looking ahead, it is foreseeable that, as more ear data becomes accessible, researchers will increasingly turn their attention towards ear biometrics as a viable means of human recognition. This nascent avenue holds significant promise for the future of biometric identification,” he tells KUST Review.

Join our mailing list

Get the latest articles, news and other updates from Khalifa University Science and Tech Review magazine