smartphones be unlocked?
People unlock their mobile phones, from the initial button and pattern unlocking to the currently widely used fingerprint unlocking, voice unlocking and face recognition. Recently, a Japanese scientific research provided another brand-new solution for unlocking smartphones - unlocking by "blowing a breath".
(Source: Kyushu University)
Kyushu University in Japan and The University of Tokyo scientific research team collaborated to analyze the compounds in human breath through an artificial olfactory sensor system for the first time, thereby achieving the effect of verifying personal identity. They used a 16-channel chemical resistance sensor array combined with machine learning techniques to create an "artificial nose" and successfully achieved an average accuracy of more than 97%. Additionally, the study demonstrates the impact of sensor number on accuracy and reproducibility.
Recently, a related paper titled "Breath odor-based individual authentication by an artificial olfactory sensor system and machine learning" was published on Chemical Communications [1] .
Figure丨Related papers (Source: Chemical Communications)
The key to unlocking smartphones is to confirm the uniqueness of people's identities through biometric technology . In the past identification methods, the conventional approach is to identify fingerprints, palmprints , voices and faces. In addition, uncommon identification methods such as acoustics and finger veins are also included.
"These technologies rely on the uniqueness of each person's body, but they are not foolproof. Physical characteristics may be copied or even damaged by injury," said the paper's first author, Department of Materials Science and Engineering, Graduate School of Engineering, University of Tokyo Ph.D. student Chaiyanut Jirayupat told the media, "Recently, human scent has emerged as a new biometric identity, mainly using your unique chemical composition to confirm who you are."
Picture | Artificial olfactory sensor for breath-based biometric authentication (Source: Kyushu University)
To achieve such an effect, the team first thought of using compounds produced from human skin to identify identity information. However, this approach is not safe. The contradiction is that the machine must meet a certain concentration to detect identity information in order to achieve "recognition", and the skin cannot produce high concentrations of volatile compounds.
Based on this, researchers came up with another idea: Can human breath be used for biometric identification?
"Concentrations of volatile compounds in the skin can be as low as parts per billion or trillion, while compounds exhaled from the breath can be as high as parts per million," Girayupat continued, "In fact, humans Breathing has already been used to identify whether a person has cancer, diabetes, or even COVID-19. (Source: Kyushu University) After identifying this new biometric approach, the team first investigated. The subjects' breath samples were comprehensively analyzed to determine which compounds could be used for biometric authentication. According to the results, the sensor can identify components of human breath, including 28 compounds, for biometric identification.
Specifically, the team built an "artificial nose" olfactory sensing system, and each of the 16 channels of the olfactory sensor array can identify a specific range of compounds. The sensor data is then transmitted to a machine learning system, which interprets the chemical composition of each breath sample.
Figure 丨 Graphical workflow for individual authentication based on breath odor perception (Source: Chemical Communications)
In related experiments, the team used the system to detect breath samples of 6 subjects. The results showed that the system was not only able to identify individuals, but achieved an average accuracy of 97.8%. The subject group was composed of people of different ages, genders and nationalities. When the researchers further expanded the sample size to 20 subjects, the accuracy rate remained at a high level and did not affect the accuracy due to the expansion of the sample size. accuracy.
Will the respiratory composition of the same person be affected by dietary factors, thus affecting the test results? With this in mind, subjects were fasted 6 hours before the relevant test to ensure test accuracy.
Of course, if this system is to be used on a large scale in the future, more technical problems will need to be solved, such as making the relevant results no longer affected by factors such as diet, but can accurately identify the use in any state through breathing. the person’s identity information.
However, the team is optimistic about solving the problem. Takeshi Yanagida, corresponding author of the paper and professor at the Graduate School of Engineering at the University of Tokyo, told the media: "Thankfully, our current research shows that adding more sensors and collecting more data can overcome this obstacle. ”
-End-
Reference:
1. Chaiyanut Jirayupat et al. Chemical Communications 58, 6377-6380 (2022) https://doi. org/10.1039/D1CC06384G
https://www.kyushu-u.ac.jp/ en/researches/view/233
