A Man Walks onto a Train
Sitting on the train, a wave of scent washes over you as a new passenger makes their way to the seat next to you. Scrunching up your nose, you scoot to the right as much as you can without being impolite. Body odor. It can make us intrigued, attracted, disgusted. We can identify familiar people almost immediately from their scent. But what is it? And how is it at the forefront of an emerging field in information security?
Individuals are thought to have their own distinctive scent, which has many evolutionary uses including communication, attracting mates, assertion of territorial rights, and protection from a predator. Body odor can be broken down into three components: the primary odor of an individual is stable and does not change over time. These may be due to genetic influences. The secondary odor contains components influenced by diet and environmental factors. The third odor contains components of outside sources, including soaps, lotions, and perfumes. On a molecular level, our odor is composed of chemicals known as volatiles which include hydrocarbons, alcohols, carboxylic acids, ketones and aldehydes. While an individual's odor has been shown to vary slightly based on menstrual cycle, emotional state, health and age[4–6], it has also been shown that individuals retain their own scent throughout their life[7,8]. The existence of a scent “cloud” surrounding our body has been proven by the National Institute for Medical Research of London. Their research shows that this cloud-which is produced from bacteria in dead skin cells- has a width of 1-4cm and contains 4x more germs than the air surrounding the cloud. This cloud is quite pervasive and probably contains and emits our unique body odor.
An Emerging Field
At the intersection of the internet of things, robust machine learning algorithms, and information security is the growing field of biometrics. While you may not have heard the term biometrics before, chances are you have used them in the past few hours. Each time you use fingerprint scanning or facial recognition to open your phone, you are using biometrics. Outside of the convenience of not having to type in a password to unlock our phone, biometrics are also used in crime scene investigation and in the medical field for identification purposes. Biometric identifiers are distinctive, measurable characteristics used to identify individuals. Common biometrics include DNA, fingerprint, face, voice, and retinal recognition. As the field of biometrics has emerged, so too have challenges to the technology. Low accuracy in the technologies continue to persist. People are hesitant to offer personal biometrics such as fingerprints and DNA. Recently, facial recognition software has been disrupted by lasers in Hong Kong by protesters seeking to avoid recognition by law enforcement. To overcome these challenges, novel biometrics that are contactless, secure, and robust are currently being explored.
Mosquitoes, Bloodhounds, and Wine
Back before facial recognition, fingerprint analysis, and DNA forensics were used in crime scene investigation, law enforcement officers used bloodhounds to identify culprits of a crime. In one study, given only a very small fragment of a bomb, trained dogs could identify the culprit 6o% of the time. This ability to recognize body odor shows potential for using body odor as a biometric identifier[14–16]. And bloodhounds aren’t alone in their ability to detect body odor- mosquitoes have been shown to be more attracted to some individuals than others depending on variation in chemical cues[17,18]!
Research into odor recognition is exploding. Odor analysis is widely used in quality control in food[19–23], tea[24,25], and wine[26–28]. Odor analysis can determine the maturity and freshness in fruits[29,30]. In the medical field, odor is being used to detect diseases including COPD, gastrointestinal toxicity, kidney disorders, lung cancer, and breast cancer[31–35].
What's That Smell?
Given the stability of personal body odor and the success of odor detection and identification, odor biometrics may be the next biometric technology used in information security. The advantages to using odor biometrics lie in its contactless approach, the strong authentication it exhibits due to the fact that it is currently impossible to replicate body odor, and its high accuracy. Contrary to biometric techniques such as facial recognition, which has a high error rate, and fingerprint technology, which requires contact, odor biometrics yields high accuracy without contact.
Preliminary studies have shown success in identifying individuals based on body odor. Using samples of odor taken from the hand, Curran et.al. evaluated 10 subjects’ body odor and determined “primary odor constituents” of each participant. Their rank system showed accurate identification in 99.54% of cases. Wongchoosuk, et. al. developed an electronic nose that, combined with a principal component analysis algorithm, could recognize individuals even after the application of deodorant. In an analysis of mass-spectrometry data for odor biometric identification, Rodriguez-Lujan et.al. achieved recognition rates over 85%. Perhaps the most convincing argument for odor as a biometric comes from Penn, et. al. In the study, 197 adults were sampled repeatedly over time, controlling for effects of a variety of potential confounding factors. Results showed not only that individuals have unique body odor compounds, but researchers could also identify gender of an individual by body odor alone.
E-T or E-Nose?
Odor biometrics combines sensing technologies, machine learning algorithms, and statistics.
The odor sensing system enables the tracing of odor from the environment. This system, affectionately known as electronic or “E” noses, can be a single sensing device (like a gas chromatograph and spectrometer) or can be an array of chemical sensors[1,2]. This electronic nose performs similarly to your own noise, sensing specific chemicals in the air.
Once the chemical smells have been sensed with the E-nose, the recognition process begins. The recognition process combines machine learning and statistical techniques to generate features. This feature-generator (commonly principal component analysis, PCA) extracts discriminatory features from the “smells” and generates a digital string of features known as a biometric template. This biometric template can then be used to classify a person based on their unique odor profile.
A Scents of the Future
Now, before you can line up to get your new iPhone with an odor-detecting unlock feature, there is still significant research still to be done. There are many challenges with using odor biometrics. Odors are very complex and validation of E-noses and their associated algorithms across diverse populations needs to occur for future development. Odor biometrics is an exciting frontier in the emerging field of biometrics for information security. With the rapid growth of the internet of things (IoT) and sensor technologies, we may be able to smell the future.
1. Wongchoosuk C, Lutz M, Kerdcharoen T. Detection and classification of human body odor using an electronic nose. Sensors. 2009;9(9):7234-7249. doi:10.3390/s90907234
2. Inbavalli P, Nandhini G. Body Odor as a Biometric Authentication. www.ijcsit.com. Accessed October 26, 2019.
3. Sastry S, Buck K, Janak J, Dressler M, Preti G. Volatiles emitted by humans. Biochem Appl Mass Spectrom. 1980:1085-1129.
4. Penn D, Potts WK. Chemical signals and parasite-mediated sexual selection. Trends Ecol Evol. 1998;13(10):391-396. doi:10.1016/S0169-5347(98)01473-6
5. Ackerl K, Atzmueller M, Grammer K. The scent of fear. Neuroendocrinol Lett. 2002;23(2):79-84.
6. Singh D, Bronstad PM. Female body odour is a potential cue to ovulation. Proc R Soc B Biol Sci. 2001;268(1469):797-801. doi:10.1098/rspb.2001.1589
7. Romanes GJ. Experiments on the Sense of Smell in Dogs. J Linn Soc London, Zool. 1887;20(117):65-70. doi:10.1111/j.1096-3642.1887.tb01437.x
8. Penn DJ, Oberzaucher E, Grammer K, et al. Individual and gender fingerprints in human body odour. J R Soc Interface. 2007;4(13):331-340. doi:10.1098/rsif.2006.0182
9. Doyle C. The secret cloud that surrounds us. Fam Heal. 1970;32.
10. Jain A, Hong L, Pankanti S. Biometric identification. Commun ACM. 2000;43(2):90-98. doi:10.1145/328236.328110
11. Biometrics and biometric data: What is it and is it secure? https://us.norton.com/internetsecurity-iot-biometrics-how-do-they-work-are-they-safe.html. Accessed October 26, 2019.
12. Hong Kong protesters use lasers to avoid facial recognition cameras and blind police | The Independent. https://www.independent.co.uk/news/world/asia/hong-kong-protests-lasers-facial-recognition-ai-china-police-a9033046.html. Accessed October 26, 2019.
13. Stockham R. FBI — Forensic Science Communications - July 2004. https://archives.fbi.gov/archives/about-us/lab/forensic-science-communications/fsc/july2004/research/2004_03_research03.htm. Accessed October 26, 2019.
14. Jenkins S. Can Police Dogs Identify Criminal Suspects By smell? In: Oxford University Press. ; 2004.
15. Schoon GAA. Scent identification lineups by dogs (Canis familiaris): Experimental design and forensic application. Appl Anim Behav Sci. 1996;49(3):257-267. doi:10.1016/0168-1591(95)00656-7
16. Rodriguez-Lujan I, Bailador G, Sanchez-Avila C, Herrero A, Vidal-De-Miguel G. Analysis of Pattern Recognition and Dimensionality Reduction Techniques for Odor Biometrics. Vol 52. doi:10.1016/j.knosys.2013.08.002
17. Qiu YT, Smallegange RC, Van Loon JJA, Ter Braak CJF, Takken W. Interindividual variation in the attractiveness of human odours to the malaria mosquito Anopheles gambiae s. s. Med Vet Entomol. 2006;20(3):280-287. doi:10.1111/j.1365-2915.2006.00627.x
18. Schreck CE, Kline DL, Carlson DA. Mosquito attraction to substances from the skin of different humans. J Am Mosq Control Assoc. 1990;6(3):406-410. http://www.ncbi.nlm.nih.gov/pubmed/2230769. Accessed October 26, 2019.
19. Barié N, Bücking M, Rapp M. A novel electronic nose based on miniaturized SAW sensor arrays coupled with SPME enhanced headspace-analysis and its use for rapid determination of volatile organic compounds in food quality monitoring. Sensors Actuators, B Chem. 2006;114(1):482-488. doi:10.1016/j.snb.2005.06.051
20. Vestergaard JS, Martens M, Turkki P. Application of an Electronic Nose System for Prediction of Sensory Quality Changes of a Meat Product (Pizza Topping) During Storage. LWT. 2007;40:1095-1101.
21. Vinaixa M, Vergara A, Duran C, Llobet E, Badia C. Fast detection of rancidity in potato crisps using e-noses based on mass spectrometry or gas sensors. Sensors Actuators B. 2005;106:67-75. doi:10.1016/j.snb.2004.05.038
22. Panigrahi S, Balasubramanian S, Gu H, Logue CM, Marchello M. Design and development of a metal oxide based electronic nose for spoilage classification of beef. Sensors Actuators, B Chem. 2006;119(1):2-14. doi:10.1016/j.snb.2005.03.120
23. El Barbri N, Llobet E, El Bari N, Correig X, Bouchikhi B. Electronic Nose Based on Metal Oxide Semiconductor Sensors as an Alternative Technique for the Spoilage Classification of Red Meat. Sensors (Basel). 2008;8(1):142-156. doi:10.3390/s8010142
24. Yu H, Wang Y, Wang J. Identification of tea storage times by linear discrimination analysis and back-propagation neural network techniques based on the eigenvalues of principal components analysis of E-nose sensor signals. Sensors. 2009;9(10):8073-8082. doi:10.3390/s91008073
25. Dutta R, Hines EL, Gardner JW, Kashwan KR, Bhuyan M. Tea quality prediction using a tin oxide-based electronic nose: An artificial intelligence approach. Sensors Actuators, B Chem. 2003;94(2):228-237. doi:10.1016/S0925-4005(03)00367-8
26. Marti MP, Busto O, Guasch J, Boque R. Electronic Noses in the Quality Control of Alcoholic Beverages. Trends. Anal Chem. 2005;24:57-66.
27. Garcia M, Aleixandre M, Gutierrez J, Horrillo MC. Electronic Nose for Wine Discrimination. Sens Actuat. 2006;113:911-916.
28. Gibbs M. Biometrics: body odor authentication perception and acceptance. ACM SIGCAS Comput Soc. 2010;40(4):16-24. doi:10.1145/1929609.1929612
29. Kinjo H, Oshiro N, Duong SC. Fruit maturity detection using neural network and an odor sensor: Toward a quick detection. In: 2015 10th Asian Control Conference: Emerging Control Techniques for a Sustainable World, ASCC 2015. Institute of Electrical and Electronics Engineers Inc.; 2015. doi:10.1109/ASCC.2015.7244428
30. Lee KM, Son M, Kang JH, et al. A triangle study of human, instrument and bioelectronic nose for non-destructive sensing of seafood freshness. Sci Rep. 2018;8(1):547. doi:10.1038/s41598-017-19033-y
31. Covington JA, Wedlake L, Andreyev J, et al. The detection of patients at risk of gastrointestinal toxicity during pelvic radiotherapy by electronic nose and FAIMS: a pilot study. Sensors (Basel). 2012;12(10):13002-13018. doi:10.3390/s121013002
32. Dymerski T, Gebicki J, Wiśniewska P, Śliwińska M, Wardencki W, Namieśnik J. Application of the electronic nose technique to differentiation between model mixtures with COPD markers. Sensors (Switzerland). 2013;13(4):5008-5027. doi:10.3390/s130405008
33. Di Natale C, Mantini A, Macagnano A, Antuzzi D, Paolesse R, D’Amico A. Electronic nose analysis of urine samples containing blood. Physiol Meas. 1999;20(4):377-384. http://www.ncbi.nlm.nih.gov/pubmed/10593231. Accessed October 26, 2019.
34. Phillips M, Cataneo RN, Ditkoff BA, et al. Volatile markers of breast cancer in the breath. Breast J. 2003;9(3):184-191. doi:10.1046/j.1524-4741.2003.09309.x
35. Phillips M, Cataneo RN, Cummin ARC, et al. Detection of lung cancer with volatile markers in the breath. Chest. 2003;123(6):2115-2123. doi:10.1378/chest.123.6.2115
36. Curran AM, Prada PA, Furton KG. The differentiation of the volatile organic signatures of individuals through SPME-GC/ms of characteristic human scent compounds. J Forensic Sci. 2010;55(1):50-57. doi:10.1111/j.1556-4029.2009.01236.x
37. C OA, M FT, S BR, A AA. An Exploratory Study Of Odor Biometrics Modality For Human Recognition. www.ijert.org. Accessed October 26, 2019.
Author and Illustrator: Brinnae Bent
Brinnae is a PhD student studying the intersection of data science, engineering, and healthcare. She can frequently be found building machine learning models, running in ultramarathons, and working on her SciComm presence.