Eventually, if someone asks to smell you after you fill out the SF-86, it may not be creepy (don’t worry – for now it’s still creepy). Last month researchers in Japan released a study based on breath-based biometrics. Thanks to machine learning an an artificial nose built with a 16 sensor array, Kyushu University’s Institute for Materials Chemistry and Engineering was able to build a biometric sensor with 97% accuracy – based on your smell.
Olfactory sensors have been tested in the past, but a variety of factors make your overall smell not the best source for a biometric scan. Breath, on the other hand, has already proven successful at identifying illnesses and health issues. The olfactory sensor proved an average accuracy of 97.8% – not far off from other popular biometric strategies like fingerprints and iris scans.
The research is currently heavily within the realm of academic research. While articles have already speculated that the technology could be used to open your smartphone, that type of hands-on application would be far into the distant future. In the current process, a machine is used to create a ‘breath profile’.
With fingerprints the gold standard in biometrics, including in the security clearance process, one might wonder why one might consider any other biometric. Fingerprints have many advantages – accuracy, easy of use – but they also have one natural downside: they are physical, and therefore may be manipulated. Comparing current biometrics like finger, palm or face, one of the study authors pushes back on why the smell test may be a viable biometric for the future.
“These techniques rely on the physical uniqueness of each individual, but they are not fool proof. Physical characteristics can be copied, or even compromised by injury,” said Chaiyanut Jirayupat, Kyushu University and study author. “Recently, human scent has been emerging as a new class of biometric authentication, essentially using your unique chemical composition to confirm who you are.”
