Research group from the University of Georgia has developed a system for recognition of the ears with an accuracy of 99%, which can become a reliable alternative technology requiring recognition of faces or fingerprints.
Like fingerprints, ears are unique to each person – even in the same -eating twins they are different, the researchers say. Unlike a person, the ears remain relatively unchanged over the years, except that the earlobe drops below over time.
According to researchers, this technology works almost in the same way as recognizing fingerprints or face prints. She scans your ears and draws their biometric picture, writing down several samples in the device. Then you can unlock the device if your ears match the logs.
“The phone captures several samples of a person’s personality, and the images are temporarily preserved on your device. By analogy with a fingerprint to unlock your phone, you can use your ear,” said the leading author of the study by Tirimachos Burlai.
Ear recognition system based on deep learning. The system previously studies at the Imageenet data set, and then sets up using a set of images of ear
Earlier this year, Apple filed a patent in which it is assumed that AirPods can be used to identify a person in the form of his ear channel and prevent their use by unauthorized persons.
The research group of the University of Georgia sees the application of its technology in situations where it is impossible to recognize persons, for example, when someone wears a mask. According to researchers, it can also be used to improve existing security systems at airports, as well as video surveillance systems.
Further studies will be focused on making an ear recognition algorithm to work with thermal images and in darker conditions, when ordinary cameras cannot get a clear image.