NEC develops biometrics technology that uses sound to distinguish individually unique ear cavity shape

NEC Corporation announced the development of a new biometric personal identification technology that uses the resonation of sound determined by the shape of human ear cavities to distinguish individuals.

The new technology instantaneously measures (within approximately one second) acoustic characteristics determined by the shape of the ear, which is unique for each person, using an earphone with a built-in microphone to collect earphone-generated sounds as they resonate within ear cavities. This unique method for extracting features is useful for distinguishing individuals based on acoustic characteristics and enables rapid and highly accurate recognition (greater than 99% accuracy).

„Since the new technology does not require particular actions such as scanning a part of the body over an authentication device, it enables a natural way of conducting continuous authentication, even during movement and while performing work, simply by wearing an earphone with a built-in microphone to listen to the sounds within ears,” said Shigeki Yamagata, general manager, Information and Media Processing Laboratories, NEC Corporation.

NEC plans to commercialize the technology within FY2018 in a wide range of applications, including the prevention of identity fraud in operations related to safety and security, such as in maintenance, management, and security of critical infrastructure, in ensuring confidentiality of wireless communications and telephone calls, and in voice guidance services designed for particular individuals or particular scenarios.

Key features of this new technology are as follows:

1. Instantaneous and stable measurement of individually unique acoustic characteristics of the human ear
An earphone with a built-in microphone is used to generate a few hundred milliseconds of acoustic signals from the earphone speaker and to receive the signals transmitted within the ear through the microphone. During this process, a synchronous addition method, which adds and obtains the average of the waveforms of the multiple signals received, is used to eliminate noise from the received signals. It then calculates how the sound resonates within the ear (acoustics). These steps are carried out instantaneously (within one second), enabling a stable and rapid means for measuring individually unique acoustic characteristics.

2. Extraction of features based on the unique structure of the human ear to achieve accurate recognition
The extraction of feature values from the acoustic characteristics is based on the knowhow from NEC’s advanced biometrics recognition technologies,*1 which have been proven to be the world’s most accurate. Since the shape and size of the ear are unique for each person, these acoustics-based features can be used in distinguishing individuals. Results of experiments have shown that, in particular, the signal components that travel through the external ear canal and are reflected by the tympanic membrane, as well as those signal components that pass through the tympanic membrane and are reflected within the inner parts of the ear, are both important for recognition. Thus, extracting feature values from these two types of signal components has enabled recognition operations with minimal computational complexity, realizing stable and highly accurate recognition (greater than 99% accuracy).