A recent study has determined that people’s identities can match their unique dance styles,media new Atlas reported. Scientists at the University of Weszley in Finland first used motion-capture techniques to see if they could determine the psychological characteristics of the subjects from their dance patterns – including their mood, empathy levels, and whether they were extroverted or neurotic.
The researchers were interested in whether it could simply determine what type of music a person’s dance would be. The accuracy of this work is only about 30%. What they have inadvertently discovered, however, is that everyone, regardless of the type of music, has a unique dance style that can be identified and particularly suited to him. This is done by combining motion capture techniques using machine learning algorithms.
In the study, 73 volunteers each danced eight kinds of music, including blues, country, dance/electric, jazz, metal, pop, reggae and rap. Participants received no instructions except “move in a natural manner.” Once the system has initially determined which person dances in which way, it can then identify them – only their dance moves – with an accuracy rate of 94%.
Although the technology is suitable for any music genre, some of them provide better results. In the case of metals, in terms of accurate IDs, it is not the largest. Scientists believe this may be because of the fact that certain actions are closely related to the style of music, so most people may choose them rather than express themselves exclusively.
Dr Emily Carlson, lead author of the study, said: “We are less interested in applications such as surveillance than in the results that tell us about human musicality. We have a lot of new questions to ask, such as whether our action characteristics remain the same throughout our life cycle, whether we can detect differences between different cultures based on these movement characteristics, and how humans are able to identify individuals by comparing their dance moves to computers. “
The paper was recently published in the journal Of New Music Research.