November 11, 2019

The world is an increasingly busy place, and scheduling that appointment with your doctor for a check up might be on the end of your to-do list. As new technologies enter into the world of health and medicine, it is becoming increasingly easier to check your own vitals to ensure you’re living the healthiest lifestyle possible.

Enter IEEE Member Yangong Zheng and the electronic nose technology being used to detect and diagnose diseases and disorders simply through the smell your body emits. Although the biological olfactory system was inspiration for this type of technology, an electronic nose does not physically look like or depict a human nose.

“Artificial systems for noninvasive chemical sensing are commonly referred to as ‘electronic nose technology’,” explains Zheng.

See also Meet the E-Nose That Actually Sniffs

How Does Electronic Nose Technology Work?

Electronic noses use chemical sensors to detect metal oxides or conducting polymers.

“The electronic nose is a multisensory system employed in the analysis of a complex gas environment that transforms raw data into specific recognition patterns for qualitative or quantitative recognition through computational and statistical analysis,” says Zheng. “Therefore, the components of electronic noses typically combine a sampling system, a sensor array and a pattern classifier.”

Pattern recognition takes the sensor data and extracts the informative features of the gas components in order for the electronic nose to detect the suspicious smells.

“The goal for pattern recognition is classified to qualitative or quantitative analysis,” says Zheng. “Qualitative analysis is to distinguish the species of gas, and quantitative analysis is to estimate the concentration of gas.”

The electronic nose technology also uses a variety of machine learning algorithms and integrated AI components for accurate detection.

“Many machine learning algorithms were widely integrated with the electronic nose for qualitative analysis, such as artificial neural networks, probabilistic neural networks, self-organizing maps, deep learning, decision trees, support vector machines and principal component analysis,” says Zheng.

See also Improving the Quality of Sleep with AI and Machine Learning

Disease and Disorder Diagnosis

While the sensor also has the ability to detect food quality, environmental pollution and indoor air monitoring, one of the most vital areas for this research has been the sensor’s ability to use human odor recognition for medical diagnosis.

“Human odor, which results from internal hormonal or metabolic changes, is found to contain a biomarker for certain diseases and reflects the physiological status of an individual,” says Zheng. “The major sources of human odor are breath, saliva, sweat, skin and urine. Recently, the components of exhaled breath, skin emissions and urine odor have attracted attention as diagnostic biomarkers of diseases and disorders, such as diabetes, liver disease, asthma, lung cancer and chronic obstructive pulmonary disease and cystic fibrosis.”

Zheng programmed the gas sensor to detect a variety of different gases that could indicate your vitals are at a concerning level in a matter of minutes.

“Generally, the gas sensor has a broad spectrum to gases,” says Zheng. “Volatile organic compounds (VOCs) are the major target for the electronic nose in the medical diagnosis, such as acetone for patients with diabetes, ammonia in individuals with liver disease, pentane in patients with asthma and alkanes in patients with lung cancer.”

While the electronic nose technology has the potential to be used for diagnosis and as an indicator of overall health, Zheng reiterates that it is is not yet mature enough to perform a full diagnosis. Once a detection indicates a warning, visit a doctor to get a full analysis of your health condition. Once we have a little more knowledge on our side, finding the motivation to finally make that doctor appointment might be a lot easier.