gb2s
-->

Odour

odor2
 What is odour identification?

 

Advances in technology have made possible the advent of sensors capturing the body odour, so non-invasive biometric techniques based on this information are being developed.

The human identification from body odour is not a new idea, bloodhounds have been used to identify people for more than a century by the police. The well-known ability of these specially trained dogs to follow tracks of a person from just an odor sample shows the viability of using the body odour as biometric identifier.

Body odor also can vary from person to person. There are recognizable patterns of each person's body odour that remain steady. Therefore, every person has his/hers own odour and this would allow his/her identification within a group of people at an accuracy rate higher than 85%.

 

The body odor production is a complex process that involve different types of sweat glands:

  • eccrine glands which are spread across the skin and regulate body temperature by cooling the skin with sweat. Eccrine sweat glands are distributed all overthe body, but they are specially concentrated in the palms, forehead and soles. The sweat produced by these glands is quite abundant, even two or four liters per hour in a normal person doing exercise.

  • apocrine glands are mainly responsible for body odour, because the sweat they produce contains high levels of protein, which bacteria find easy to break down. They are only present in the armpit, pubic area and areola.

  • sebaceous glands Sebaceous glands are concentrated in body parts with hair such as breast, back or head. The secretions of these glands are commonly named sebum and they are rich in lipids.

 

 How does body odour analysis work?

 

  1. Hand odor capture: The person places the hand into a device in which air flows collecting the volatile compounds of the odor.

  2. Identification system: The acquisition system provides a mass spectrogram for each session. This spectrogram reflects the number of ions detected for each mass during the time each subject was placing the hand in the sensor and every compound can be identified with a peak in the spectrogram. This process includes two steps: feature selection, that select the most relevant subset of features from the original set of variables, and feature extraction, that allows ignoring those compounds with a low variance and therefore, not helpful to discriminate between different subjects. Last step is to use a classifier in order to discriminate the body odor of each subject.

 

  Applications

 

  • Biometric identification based on human body odour. Every individual has a specific odour that make it possible the identification among the population.

  • Detect colorectal cancer from the analysis of human breath.

  • Detect leukemia from the analysis of volatiles present in the blood at room temperature.

 

  Our Projects

 

 

Category: /

Grup of Biometrics, Biosignals and Security (GB2S)