Scientists at Universidad Politecnica de Madrid in Spain and colleagues found that there are recognisable patterns of each person's body odour that remain steady.
Therefore, every person has their own odour and this would allow their identification within a group of people at an accuracy rate of 85 per cent, researchers said.
The findings lead the way to improve personal identification that is less aggressive than other biometric techniques being used today.
Several biometric techniques like the iris recognition and fingerprint recognition have a low error rate.
However, these two techniques are usually related to criminal records and for this reason when people are required to identify themselves, they are not so willing to collaborate, researchers said.
On the other hand, other recognised biometric techniques like the face recognition have a high error rate.
The development of new sensors that allow the capture of body odour can provide a less aggressive solution, they said.
The ability of police dogs to follow the trail of a person from a sample of their personal odour is well known and proves that using body odour is an effective biometric identifier, researchers said.
Although the sensors used today have not yet achieved the accuracy dog's sense of smell, the researchers used a system developed by the Spanish company Ili Sistemas SL that has a high sensitivity to detect volatile elements present in body odour.
They analysed a group of 13 people during 28 sessions and found that recognisable patterns on each person's body odour have an identification at an error rate of 15 per cent.