November 28, 2014 -
A University of Adelaide PhD student has discovered a new process known as “body recognition” which could potentially become a new method of biometric identification in the future, according to a report by The Lead.
Teghan Lucas has been using a 1988 database of 4,000 US defence force personnel as the basis for the process, testing her body measurements against more traditional biometric forms such as facial recognition.
Lucas said that using a combination of eight or more body measurements enabled her to lower the probability of finding the same measurements to one in a quintillion.
Body recognition would be particularly useful in cases of criminal activity or missing individuals, as its accuracy does not depend on as many data points as facial recognition.
Facial recognition has become a popular method of identification, but there are some notable drawbacks to the method when a suspect’s face is covered during a criminal act, or the surveillance footage is of a low quality.
Another major issue is that facial expressions can impact perceptions and measurements of facial features.
In comparison, the larger measurements of the body are far easier to view and quantify than the small details of someone’s face.
“If somebody comes back twenty years later, claiming to be that person, as long as they’re an adult, you can compare the two to determine whether or not it’s the same person,” said Lucas, adding that an adult’s skeleton does not alter in size for the majority of their lifetime.
In addition, clothing does not impact body measurements because the gravitational force acting on the clothing will ensure that the overall shape of the body is still clear.
Additionally, any body movement captured on a surveillance video can only improve the ability of viewing the individual’s outline.
Lucas is also hoping to partner with a software developer to emulate the functionality of the proprietary Defence Force software she has been using to mark points on a body and calculate the exact distances between them.