How do body odor recognition biometrics work?

A body odor recognition system uses an electronic nose (E-nose) with numerous chemical sensors to detect and identify volatile organic compounds (VOCs) in the air. General-purpose E-noses are available to detect and identify complex chemical mixtures that constitute aromas, odors, fragrances, and formulations. Breathalyzers are specialized E-noses used to measure the amount of alcohol in a person’s breath.

An experimental E-nose has been developed to detect and classify human armpit body odor and distinguish between individuals. This article begins with a general overview of E-nose operation, looks at how E-noses can be applied to identifying human body odors, including armpit odor, and closes by presenting an existing E-nose designed for use in industrial settings.

E-noses are designed to mimic the functionalities of a biological nose (Figure 1). They use sensors to generate signals analogous to the olfactory epithelium receptors in a biological nose. A microcontroller unit (MCU) in an E-nose replaces the signal processing function of the olfactory bulb. From there, the signals are sent to a pattern recognition engine for identification using a database of odor signatures and some designs for continuous learning.

Figure 1. Comparative functionalities of a biological nose and an E-nose. (Image: MDPI sensors)

Body odors

Studies have demonstrated that each person has a unique body odor signature that carries information about their genetic makeup and personal health and environmental variables, such as diet and hygiene. That unique odor can potentially be separated from other odors caused by soaps, perfumes, and other factors.

A body odor database, like fingerprints and facial recognition, would have to be developed for it to be useful. Otherwise, body odor recognition is limited to distinguishing between individuals, not identifying or authenticating specific individuals.

In one case, an E-nose has been fabricated, employing an array of metal oxide sensors to identify human armpit odors. The measurement circuit uses a voltage divider resistor network to measure the output of each sensor. Custom software controls the E-nose through a USB data acquisition card. A principal component analysis (PCA) algorithm implements pattern recognition and classification.

The software also compensates for the effect of humidity on the odor analysis. The resulting E-nose system successfully recognizes individuals even after applying various deodorants.

Industrial odors and neural networks

A fully integrated handheld E-nose chemical vapor sensing instrument has been developed for industrial applications. It uses nanocomposite sensors and pattern recognition algorithms to detect and classify the sampled chemical vapor or mixture (Figure 2).

This industrial E-nose can identify mixtures or individual chemical compounds. It’s designed for food and beverage, packaging materials, petrochemical and chemical plants, plastics, pulp and paper operations, and medical research.

This LCD screen can display the results as simple messages like “Accept,” “Reject,” “Contaminated,” and “Identified.” It also assesses the statistical quality of the results on a scale of 1 to 5. The integrated software includes a “learn” function that enables users to develop odor libraries for specific products and applications.

Summary

An odor-recognition device operates similarly to a biological nose. E-noses are currently available for industrial applications and have been demonstrated to identify and authenticate people. However, more development, including extensive and robust odor profile databases, is required before odor recognition can become a practical tool for identifying individuals in real-world settings and applications.