Body odor as a biometric identifier
The research of the Group of Biometrics, Biosignals and Security (GB2S) of the Universidad Politécnica de Madrid (UPM) in collaboration with Ilía Sistemas SL unveils that there are recognizable patterns of each person's body odor that remain steady. Therefore, every person has his/hers own odor and this would allow his/her identification within a group of people at an accurate rate higher than 85%. This result leads the way to improve personal identification that is less aggressive than other biometric techniques being used today.
Nowadays, our identity verification in most airports or border checkpoints is based on our physical resemblance to our ID card or passport photo. Even though the new electronic passports are difficult to forge, the usage of biometric techniques based on a person's physical features would increase the safety effectiveness of the border control checkpoints.
Several biometric techniques like the iris and fingerprint have a low error rate. However, these two techniques are usually related to criminal records and for this reason when a person is required to identify himself is reticent to collaborate. On the other hand, other recognized biometric techniques like the face recognition have a high error rate. Therefore, the development of new sensors that allow the capture of body odor can provide a less aggressive solution because the identification could be at the same time when crossing the system stall.
People body odor identification is not a new idea considering since it has been conducting for over a century by the police force thanks to the help of bloodhounds dogs which are trained for such task. The ability of these dogs to follow the trail of a person from a sample of his or hers personal odor is well known and proofs that using body odor is effective is an effective biometric identifier. Although the sensors used today have not yet achieved the accuracy dog's sense of smell, the research has used a system developed by the Ilí Sistemas SL company that has a high sensitivity to detect volatile elements present in body odor.
Body odor can vary considerably by actions such as diseases, diet changes or even the mood swings. However, the research carried out by the group of the GB2S of the UPM showed that, the analysis of a group of 13 people during 28 sessions have proven that recognizable patterns on each person body odor have an identification at rate error of 15%. This successful experiment and result confirms that odor biometric system as a novel technique with an enormous potential.
Although this research was conducted within the framework of Emoción Proyect which is focused on citizen safety, body odor analysis can be used in many other fields. GB2S group along with the Hospital Infanta Sofía and the Ilía Sistemas SL and SEADM SL are in fact collaborating these days on particular official projects that search for blood and breathe characteristics that can detect early signs of colon cancer and leukemia.
RODRÍGUEZ-LUJÁN, I.; BAILADOR, G.; SÁNCHEZ-ÁVILA, C.; HERRERO, A; VIDAL DE MIGUEL, G. "Analysis of pattern recognition and dimensionality reduction techniques for odor biometrics". Knowledge-Based Systems 52 (279-289). DOI: 10.1016/j.knosys.2013.08.002. Nov. 2013.
Provided by Universidad Politécnica de Madrid