When parents explain the world to their children, they do so in a way suited to the child’s age. Researchers now say: a computer or smartphone could be able to do just the same.
By Marcin Bugdol
Age estimation is an important factor in establishing a person’s identity. However, a simple division into the categories of adult vs. underage, or even a more precise division into baby vs. child vs. teenager vs. adult, is often not enough.
As adolescence results in changes to children’s voices, the relationship between a person’s age and their voice has become the subject of scientific investigations. A research team of biomedical engineers and anthropologists from Poland recently developed a method for assessing the age of children and adolescents based on a few vowels spoken.
Their proposed technique, explained in a recent article in the journal Biomedical Engineering, has a broad spectrum of potential applications, especially in cases where only a speech signal is available and there is a need to determine the speaker’s age. Automatic Speech Recognition Systems, which are already widely used, could adjust their actions to the user’s age, allowing customization without external (in particular, parental) supervision. This could, for example, improve the quality of human-robot interactions.
“This innovative method of voice-based assessment of age”, Dr. Marcin Bugdol, an author of the study, says, “may be especially useful for the police, who could, on the basis of recorded conversations, assess whether they are dealing with a minor, and if so, how old he or she is. Such recordings could come from monitoring devices, recorded telephone conversations or mounted wiretaps.”
For the algorithm to determine a person’s age, it only needs to hear a few basic vowels, such as “a” or “e”. Information about a person’s height, available for example from surveillance cameras, reduces the estimation error to about 2-3 months above or below the person’s real age. All that is needed to use the algorithm is a computer or mobile phone with a microphone and internet access, and it works regardless of where someone lives, what language they speak and how they look.
To find out more, read the original article here: