A team of researchers from the Center for Interdisciplinary Music Research at the University of Jyväskylä in Finland has discovered, through analysis using motion capture, that people dance in a unique, distinctive way, so distinctive that an artificial intelligence can recognize it and thus link it to an identity.

It was accidental, as Emily Carlson, the first author of the study, explains. Initially the researchers wanted to study something different, i.e. they wanted to understand if it is possible to use the technique of automatic learning, a well-known artificial intelligence algorithm, to identify what kind of music a dancing subject is listening to. Based on the movements, according to the researchers, perhaps it could be possible to understand the musical genre for which the subject is following the rhythm.

The experiments were conducted on 73 participants using the motion capture technique. The participants were played various pieces of music from different genres. The participants themselves had to move and dance, trying to follow the rhythm as naturally as possible.

By analyzing the participants’ movements using the machine learning technique, the algorithm was only able to identify the correct musical genre in 30% of cases.

However, the researchers themselves discovered something that is perhaps even more important: the computer could identify which of the 73 subjects was dancing 94% of the time.

“It seems that a person’s dance movements are a kind of fingerprint,” says Pasi Saari, one of the authors of the study. “Each person has a unique movement signature that remains unchanged, regardless of the type of music they are playing.”

For some genres, however, the computer was more precise than others.

It is still too early to say whether a dance recognition software can be used in the same way as a “trivial” facial recognition software, but the researchers themselves already state at this point that they are little interested in any applications in the context of surveillance or trivial recognition.